Harnessing Machine Learning to Analyze Renewable Energy Research in Latin America and the Caribbean

Abstract

The transition to renewable energy is essential for mitigating climate change and promoting sustainable development, particularly in Latin America and the Caribbean (LAC). Despite its vast potential, the region faces structural and economic challenges that hinder a sustainable energy transition. Understanding scientific production in this field is key to shaping policy, investment, and technological progress. The primary objective of this study is to conduct a large-scale, data-driven analysis of renewable energy research in LAC, mapping its thematic evolution, collaboration networks, and key research trends over the past three decades. To achieve this, machine learning-based topic modeling and network analysis were applied to examine research trends in renewable energy in LAC. A dataset of 18,780 publications (1994–2024) from Scopus and Web of Science was analyzed using Latent Dirichlet Allocation (LDA) to uncover thematic structures. Network analysis assessed collaboration patterns and regional integration in research. Findings indicate a growing focus on solar, wind, and bioenergy advancements, alongside increasing attention to climate change policies, energy storage, and microgrid optimization. Artificial intelligence (AI) applications in energy management are emerging, mirroring global trends. However, research disparities persist, with Brazil, Mexico, and Chile leading output while smaller nations remain underrepresented. International collaborations, especially with North America and Europe, play a crucial role in research development. Renewable energy research supports Sustainable Development Goals (SDGs) 7 (Affordable and Clean Energy) and 13 (Climate Action). Despite progress, challenges remain in translating research into policy and addressing governance, financing, and socio-environmental factors. AI-driven analytics offer opportunities for improved energy planning. Strengthening regional collaboration, increasing research investment, and integrating AI into policy frameworks will be crucial for advancing the energy transition in LAC. This study provides evidence-based insights for policymakers, researchers, and industry leaders.

1. Introduction

The growing global energy demand and the urgent need to reduce greenhouse gas emissions have accelerated the transition from fossil fuels to renewable energy sources as a sustainable alternative [1]. In this context, the development of renewable energy is essential for achieving a low-carbon future and fulfilling global climate commitments, particularly those outlined in the Paris Agreement and the Sustainable Development Goals (SDGs), including SDG 7 (Affordable and Clean Energy) and SDG 13 (Climate Action) [2].

Despite the global momentum, structured analyses of renewable energy research in specific regions, such as Latin America and the Caribbean (LAC), remain scarce. LAC has substantial renewable energy potential, particularly in biomass, wind, solar, and hydroelectric sources, positioning it as a key player in the global energy transition [3]. However, the region faces significant barriers, including socioeconomic inequalities, limited investment in research and development (R&D), regulatory constraints, and geopolitical challenges, necessitating a just and equitable transition approach [4,5]. Understanding scientific trends in this field is crucial for developing evidence-based policies that guide sustainable energy transitions in LAC.

Several additional obstacles hinder the large-scale implementation of renewable energy in the region. Geographic and climatic conditions limit project feasibility in some countries [6], while high initial costs, market constraints, and technical challenges related to energy storage and grid integration present further obstacles [7]. Additionally, bureaucratic and regulatory hurdles contribute to lengthy project approval times, delaying the widespread adoption of renewable technologies [8].

To address the operational challenges of renewable energy systems, particularly the intermittency and variability of solar and wind power, accurate forecasting has become indispensable for ensuring grid stability, planning capacity, and managing energy distribution efficiently. Traditional forecasting models, however, often struggle to capture the complex, non-linear relationships inherent in meteorological and renewable energy data [9]. In response, machine learning (ML) and deep learning (DL) have emerged as powerful data-driven approaches. These techniques can handle high-dimensional datasets, learn from historical patterns, and substantially improve prediction accuracy, making them a cornerstone of modern renewable energy research [10]. As emphasized by Ajibade et al. [11], ML applications in renewable energy are multidisciplinary, spanning from performance optimization of renewable energy technologies to the intelligent management of smart grids, underscoring their critical role in advancing sustainable energy systems. Similarly, Benti et al. [12] showed that ML and DL models outperform conventional methods in handling nonlinear and high-dimensional data, particularly in forecasting solar and wind power, which remain highly variable and uncertain. Lai et al. [13] provided a comprehensive survey demonstrating that ML approaches not only enhance short- and long-term forecasting but also facilitate resource allocation and grid integration of intermittent sources. Krechowicz [14] highlighted that ML-based approaches are particularly effective in predicting electricity production, offering practical tools for balancing supply and demand. Additionally, Ledmaoui et al. [10] demonstrated that AI-driven solar energy forecasting improves grid stability and supports decentralized system deployment, while Izanloo et al. [15] emphasized ML’s potential in investment decision-making for renewable projects. The integration of hybrid approaches, such as combining support vector machines with physical models or leveraging ensemble neural networks, further enhances prediction accuracy and grid reliability, directly supporting the goals of the energy transition.

Beyond forecasting, ML applications in renewable energy extend to a broad range of tasks, including investment risk assessment, the design of advanced energy materials, and the operational management of smart grids [13,15]. The field has seen the development of diverse computational models, ranging from single algorithms like Support Vector Machines (SVM) and Artificial Neural Networks (ANN) to complex hybrid and ensemble methods designed to further improve performance. This rapid growth and methodological diversification have generated a vast and intricate body of literature. Therefore, systematic analysis of this scholarly output is crucial for identifying dominant research trends, thematic structures, and persistent knowledge gaps that can guide both future research and evidence-based policymaking.

As renewable energy research in LAC expands, there is a growing need to analyze emerging trends, research priorities, and thematic developments. However, conventional literature reviews demand significant time and effort and often fail to capture the dynamic and interdisciplinary nature of renewable energy studies. To address these limitations, machine learning techniques such as Latent Dirichlet Allocation (LDA) offer a structured approach to mapping scientific knowledge [16,17].

LDA is a probabilistic modeling technique that identifies latent thematic patterns within large document collections, providing an automated and scalable alternative to manual reviews. This method has been successfully applied in various fields, including medicine and healthcare research [18,19], as well as environmental and agricultural sciences [20,21]. However, despite its wide applicability, LDA has been underutilized in renewable energy research in LAC.

By integrating LDA topic modeling with advanced bibliometric techniques, this study provides a comprehensive, data-driven analysis of research trends, thematic structures, and policy implications for renewable energy development in LAC. While previous studies have used topic modeling to explore the relationship between renewable energy and the SDGs, such as the global-scale analysis by Wang et al. [22], few have focused on the specific context of LAC. This study addresses that gap by offering a regional perspective, providing deeper insights into how renewable energy research has evolved in LAC and the key factors shaping its development.

Beyond its regional focus, this study introduces a more comprehensive methodological approach. Unlike previous research that relies solely on LDA, this analysis integrates Multidimensional Scaling (MDS) and HJ-Biplot visualization techniques [23], enhancing the interpretation of relationships between research topics. By combining these methods, this study provides a structured and detailed view of how different areas of renewable energy research in LAC interconnect and evolve over time.

Another key contribution is the temporal analysis of scientific production, which identifies long-term research trends and shifts in priorities. Understanding how renewable energy research in LAC has evolved over the past three decades offers valuable insights into the region’s changing energy landscape.

This study aims to provide a comprehensive analysis of renewable energy research in LAC, focusing on its evolution and key trends. Specifically, it seeks to achieve the following:

• Examine the evolution of renewable energy research in LAC over time.
• Identify the most prominent research topics, highlighting dominant areas of study and emerging trends shaping the scientific landscape.
• Provide evidence-based insights for policymakers, researchers, and industry stakeholders, supporting the development of strategies and policies that promote the expansion and integration of renewable energy in the region.

The remainder of this paper is organized as follows. Section 2 describes the methodology, including the data collection from Scopus and Web of Science and the application of LDA, network analysis, and multivariate visualization techniques. Section 3 presents the results, beginning with descriptive statistics of the scientific production, followed by the 18 topics identified through LDA, their temporal evolution, and the structure of regional and international collaboration networks. Section 4 discusses the findings, highlighting emerging research trends, structural disparities, and their implications for policy, investment, and regional cooperation. Finally, Section 5 concludes this study, summarizing the main contributions, reflecting on limitations, and outlining directions for future research.

2. Methodology

2.1. Information Sources, Search Strategy, and Data Collection

This study is based on the analysis of scientific documents retrieved from two widely recognized academic databases: Scopus and the Web of Science (WoS) Core Collection. These databases were selected due to their extensive coverage of high-impact scientific publications across multiple disciplines [24]. The document retrieval process employed the advanced search features of both platforms, incorporating logical operators (AND, OR) and carefully selected keywords relevant to renewable energy research in LAC, as detailed in

Table 1. Bibliographic databases and Keywords.

Bibliographic Database	Search Data	Search String	Results
Scopus	2 December 2024	TITLE-ABS-KEY ((“solar energy” OR “solar power” OR “photovoltaic”) OR (“wind energy” OR “wind power” OR “wind turbine”) OR (“hydroelectric energy” OR “hydropower” OR “hydroelectric power”) OR (“biomass energy” OR “biomass power” OR “bioenergy”)) AND AFFILCOUNTRY (“Argentina” OR “Bolivia” OR “Brazil” OR “Chile” OR “Colombia” OR “Ecuador” OR “Guyana” OR “Paraguay” OR “Peru” OR “Suriname” OR “Uruguay” OR “Venezuela” OR “Belize” OR “Costa Rica” OR “El Salvador” OR “Guatemala” OR “Honduras” OR “Nicaragua” OR “Panama” OR “Antigua and Barbuda” OR “Bahamas” OR “Barbados” OR “Cuba” OR “Dominica” OR “Grenada” OR “Haiti” OR “Jamaica” OR “Dominican Republic” OR “Saint Kitts and Nevis” OR “Saint Lucia” OR “Saint Vincent and the Grenadines” OR “Trinidad and Tobago” OR “Mexico”) AND PUBYEAR > 1993 AND PUBYEAR < 2025 AND (LIMIT-TO (DOCTYPE, “ar”) OR LIMIT-TO (DOCTYPE, “re”)) AND (LIMIT-TO (SRCTYPE, “j”))	N = 16,634
Web of Science	2 December 2024	TS = (“solar energy” OR “solar power” OR “photovoltaic” OR “wind energy” OR “wind power” OR “wind turbine” OR “hydroelectric energy” OR “hydropower” OR “hydroelectric power” OR “biomass energy” OR “biomass power” OR “bioenergy”) AND CU=(“Argentina” OR “Bolivia” OR “Brazil” OR “Chile” OR “Colombia” OR “Ecuador” OR “Guyana” OR “Paraguay” OR “Peru” OR “Suriname” OR “Uruguay” OR “Venezuela” OR “Belize” OR “Costa Rica” OR “El Salvador” OR “Guatemala” OR “Honduras” OR “Nicaragua” OR “Panama” OR “Antigua and Barbuda” OR “Bahamas” OR “Barbados” OR “Cuba” OR “Dominica” OR “Grenada” OR “Haiti” OR “Jamaica” OR “Dominican Republic” OR “Saint Kitts and Nevis” OR “Saint Lucia” OR “Saint Vincent and the Grenadines” OR “Trinidad and Tobago” OR “Mexico”)	N = 13,486

.

To ensure a regional focus, an additional institutional affiliation filter was applied, restricting results to publications in which at least one author was affiliated with an institution in LAC. This was achieved using the AFFILCOUNTRY parameter in Scopus and its equivalent in WoS.

This study considered publications spanning the last three decades (1994–2024) to provide a comprehensive historical perspective on research trends in the region. The search queries targeted primary renewable energy sources, including solar energy, wind power, hydroelectric energy, and biomass-based bioenergy, capturing a broad spectrum of technological and policy-related research. The search results yielded 16,634 documents from Scopus and 13,486 from Web of Science, forming the initial dataset for analysis.

Integrating datasets from multiple sources presents inherent challenges due to differences in article metadata formats between Scopus and Web of Science. To ensure data consistency and standardization, the bibliometrix package in R [25] was used, a widely adopted tool for bibliometric analysis [26,27,28,29,30,31]. This package facilitated automatic merging and deduplication of records, allowing for a streamlined dataset preparation process. Following this, a data cleaning protocol was implemented, which included the removal of 11,158 duplicate entries, 164 records without abstracts, 14 records lacking author affiliation data, and 4 articles with incorrect publication dates (e.g., set for 2025). After this refinement, a final dataset comprising 18,780 unique documents was consolidated, ensuring the reliability of subsequent analyses.

Although this study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol [32], it does not qualify as a systematic review in the strictest sense. Instead, the PRISMA framework was employed to enhance the transparency, rigor, and reproducibility of the literature selection and data processing stages, thereby strengthening the validity of the findings.

2.2. Descriptive Analysis of Collected Data

A preliminary descriptive analysis was conducted to examine general publication trends in renewable energy research across LAC. The analysis focused on the evolution of scientific output over time, identification of leading journals, and country-level contributions to research productivity. The total number of documents, their distribution over the years, and annual growth rates were assessed to quantify research activity dynamics. Additionally, international collaborations and co-authorship trends were examined, providing insight into the degree of integration between researchers in LAC and global scientific networks.

2.3. Topic Modeling and Latent Dirichlet Allocation

Topic modeling is an unsupervised machine learning technique designed to automatically identify the topic or topics present within a single document or a collection of documents. This method is particularly useful for uncovering latent or hidden topics that are not explicitly mentioned in the texts. These topics are represented by clusters of words that frequently co-occur in similar linguistic contexts, reflecting underlying concepts or themes [32].

In this study, the Latent Dirichlet Allocation (LDA) method was applied to identify these latent topics. LDA, a widely used unsupervised text mining technique, is based on Bayesian probabilistic models and is an extension of Probabilistic Latent Semantic Analysis (PLSA) [16,33]. The LDA model assumes that documents within a corpus are mixtures of latent topics, and each topic is defined by a specific probability distribution of words.

The fundamental principle of the LDA model is that each document is composed of a random mixture of topics, while each topic is characterized by a distribution of words that make up the vocabulary. For instance, a document might be represented as 60% related to health and 40% to other themes, illustrating the topic probability distribution for that document. LDA also assumes a fixed number of topics or categories across the entire document collection, with each term assigned a probability of belonging to a specific topic.

This probabilistic approach provides a comprehensive framework for analyzing large text corpora, allowing for the extraction of meaningful patterns and insights from unstructured textual data.

The topic identification process utilizing Latent Dirichlet Allocation (LDA) involved three main stages: (1) Preprocessing, (2) LDA Model Construction, and (3) Topic Labeling. The first two stages were executed using LDAShiny [34], an open-source R package [35], version 4.4.2, with a user-friendly web interface designed for scientific literature reviews, which integrates Bayesian methods for LDA and machine learning algorithms.

Text preprocessing was essential for ensuring consistency and removing noise from the dataset. This step involved:

• Converting all text to lowercase.
• Removing punctuation, dashes, numbers, and other non-relevant characters.
• Eliminating “stop-words” (e.g., articles, prepositions, and other common terms that do not contribute to topic discrimination).

The outcome of this process was a document-term matrix where each document was represented as a vector of word frequencies, providing the foundation for topic modeling.

The construction of the LDA model assumes that documents share topics in varying proportions, sampled from a Dirichlet distribution. To determine the optimal number of topics (k):

• Models were constructed with k ranging from 5 to 40 topics. Each model was generated using 1000 Gibbs sampling iterations [36] and the default Dirichlet parameters (α and β) provided by the topic models R package.
• The CV coherence measure [37], based on the distributional hypothesis, was employed to evaluate the interpretability and coherence of the topics, enabling the selection of the most meaningful model. The model with k = 18 achieved the highest coherence score, indicating strong internal semantic alignment. However, coherence alone was not the sole selection criterion. A key consideration in topic modeling is the trade-off between granularity and interpretability. Models with a small number of topics (k < 10) produced overly broad clusters that merged distinct research domains, for instance, conflating all solar-related research into a single undifferentiated topic. Such broad categories obscure thematic nuances and reduce both the analytical and policy relevance of the results. Conversely, models with a large number of topics (k > 30) generated fragmented and redundant clusters, which were difficult to label consistently and diminished the utility of the analysis for synthesizing broader research trends.

Since LDA does not inherently assign semantic labels to topics, a manual labeling process was conducted by an expert:

• The most frequent words associated with each topic were reviewed.
• Representative documents classified by the algorithm were analyzed to contextualize the topics.

This manual process ensured accurate, meaningful, and domain-relevant interpretation of the topics [38]. This multi-stage approach enhanced the reliability and interpretability of the topic modeling results, providing a robust framework for identifying and understanding latent themes in the dataset.

2.4. Quantitative Indices for Analyzing Topics and Trends

Given the large volume of data, a set of quantitative indices was used to analyze topic prevalence and evolution over time. Following the approach proposed by Xiong et al. [39] document-topic and topic-word distributions were aggregated to derive meaningful indicators of research trajectories.

To assess whether research topics were increasing or declining, a linear regression model was applied, using the year of publication as the independent variable and the proportion of each topic per year as the dependent variable, following the approach proposed by Griffiths and Steyvers [40]. Topics with statistically significant positive slopes (p < 0.05) were classified as emerging, whereas those with negative slopes were considered declining research areas. Non-significant slopes were interpreted as fluctuating or stable trends.

2.5. Visualization of Topics Using Multivariate Techniques

To complement the LDA topic modeling, we employed multivariate visualization methods that enable a clearer interpretation of the thematic relationships. Specifically, two complementary techniques, Multidimensional Scaling (MDS) and HJ-Biplot, were used to project the topics into a reduced-dimensional space, facilitating the identification of structural patterns and thematic proximities.

2.5.1. Multidimensional Scaling

To visualize the topics generated by the LDA model, it was used an interactive tool from the LDAvis package [41] designed to facilitate the exploration and interpretation of topic modeling results. This tool applies multidimensional scaling (MDS) to create a two-dimensional representation of topics, where the proximity of nodes reflects thematic similarities based on the word distributions obtained from the LDA model. In this visualization, the size or area of each node represents the relative prevalence of the topic within the corpus, while the distances between nodes indicate thematic relationships. Shorter distances suggest closer thematic connections, whereas greater distances point to more divergent term distributions. To quantify the degree of dissimilarity between topics, the Jensen-Shannon divergence [42] was used, a metric that measures differences in the probability distributions of terms across topics. This approach provides an intuitive framework for understanding the thematic structure and relationships within the corpus.

2.5.2. HJ-Biplot

The HJ-Biplot technique, an evolution of traditional biplots introduced by Gabriel [43], was employed to represent the relationships between topics and their associated variables in a shared low-dimensional space. Unlike earlier methods such as the GH-Biplot, which focuses on analyzing relationships among variables, or the JK-Biplot, which highlights similarities among individuals, the HJ-Biplot [36] achieves a balanced optimization of both aspects. This method combines principles from Multidimensional Scaling, Correspondence Analysis, and Factor Analysis to provide comprehensive visualization. In the resulting representation, the distances between row markers reflect the degree of similarity among observations, while the lengths of the column markers, or vectors, approximate the variability in the associated variables. The angles between column vectors offer insights into correlations; for example, acute angles indicate a strong positive correlation, obtuse angles suggest a negative correlation, and right angles denote an absence of correlation.

The analysis was conducted using Multbiplot software [44], version 18.0312, which allowed for a detailed exploration of the multivariate structure of the data. This approach enabled a nuanced understanding of the relationships between topics and variables, offering a powerful tool for uncovering latent patterns and trends within the corpus.

2.6. Collaboration Networks

Scientific collaboration is a critical driver of knowledge production and dissemination, particularly in multidisciplinary fields such as renewable energy. To assess the structure of research networks in LAC, bibliometric and network analysis techniques were applied to the dataset.

Using the bibliometrix package in R, key network centrality metrics were computed, including betweenness centrality (which identifies key knowledge brokers), closeness centrality (which measures how efficiently a country is connected within the network), and PageRank (which identifies the most influential research nodes).

Based on connectivity patterns and publication volume, countries were classified into four clusters: (1) highly central nations (e.g., Brazil, USA, UK, Germany), (2) regionally influential countries (e.g., Mexico, Chile, Colombia, Spain), (3) emerging contributors (e.g., China, India, Peru), and (4) low-participation countries (e.g., Guatemala, Honduras, El Salvador).

3. Results

This section presents the main findings of this study. It begins with a descriptive overview of scientific production in renewable energy research across Latin America and the Caribbean. The subsequent analysis reports the results of the LDA topic modeling, identifying 18 key thematic areas, and examines their temporal evolution. Multidimensional Scaling (MDS) and HJ-Biplot visualizations are then employed to analyze the structural relationships among topics and their historical development. The section concludes with an examination of collaboration patterns at national, regional, and international levels, highlighting both strengths and inequalities in the research network.

3.1. Descriptive Analysis

The scientific production on renewable energies in LAC between 1994 and 2024 exhibits a remarkable upward trend, as evidenced in

Table 2. Descriptive Analysis of Scientific Output on Renewable Energies in Latin America and the Caribbean (1994–2024).

Description	Results
MAIN INFORMATION ABOUT DATA
Timespan	1994:2024
Sources (Journals)	3193
Documents	18,780
Annual Growth Rate %	15.61
Document Average Age	6.13
Average citations per doc	20.94
AUTHORS
Authors	43,548
Authors of single-authored docs	439
AUTHORS COLLABORATION
Single-authored docs	558
Co-Authors per Doc	5.12
International co-authorships %	30.49
DOCUMENT TYPES
article	17,505
review	1275

and Figure 1. Over this period, research output has grown at an annual rate of 15.61%, accumulating a total of 18,780 documents distributed across 3193 scientific sources. This sustained increase underscores the region’s growing engagement in renewable energy research and highlights its contribution to global knowledge production in sustainable energy.

Collaboration has played an important role in this research landscape. While 558 documents were authored individually, the vast majority resulted from collaborative efforts, with an average of 5.12 co-authors per publication. Furthermore, 30.49% of the publications involved international collaborations, reflecting the increasing interconnectedness of LAC researchers with global scientific networks. These findings underscore the importance of cooperative research frameworks in fostering technological advancements and policy insights in renewable energy.

Regarding the academic impact of the publications, the dataset exhibits an average citation count of 20.94 per document, with an average article age of 6.13 years. This citation rate indicates that research in this domain maintains considerable academic relevance and influence. The vast majority of contributions are research articles (17,505), complemented by review papers (1275), suggesting a strong emphasis on generating new knowledge while also synthesizing existing research to guide future studies.

Figure 1 illustrates the annual growth of renewable energy research publications in LAC. The data reveals slow initial growth from the mid-1990s to the early 2000s, followed by a noticeable expansion from 2010 onward. A sharp increase is observed after 2015, coinciding with the Paris Agreement and global renewable energy policies, which likely stimulated research funding and international collaborations. Despite a slight fluctuation in recent years, the overall trend remains upward, indicating sustained interest and investment in this research domain.

The descriptive analysis reveals an evident inflection point in renewable energy research in LAC around 2015, as illustrated by the sharp increase in scientific output (Figure 1). This emergence does not occur in isolation; it strongly correlates with the adoption of the Paris Agreement and the launch of the United Nations’ Sustainable Development Goals (SDGs) in the same year. These global frameworks generated unprecedented momentum by stimulating research funding, encouraging international collaborations, and compelling national governments in the LAC region to align their energy policies with climate and sustainability targets. Although our study does not establish direct causality, the observed publication growth can be interpreted as a reflection of the global scientific community’s response to this policy-driven urgency surrounding climate action and energy transition.

The geographical distribution of research output, as depicted in Figure 2, indicates that Brazil leads scientific production in renewable energies, with 7261 publications, accounting for nearly 39% of the total output in LAC. Brazil is followed by Mexico (2547 articles), Colombia (1191), and Chile (1124), reflecting their significant investments in renewable energy research and technological development. Other notable contributors include Argentina (719 articles), Ecuador (283), Peru (162), Costa Rica (113), Cuba (97), and Uruguay (76), demonstrating a strong regional interest in sustainable energy solutions.

Conversely, smaller research outputs were recorded in some countries, including Paraguay (12 articles), Bolivia (11), Honduras (10), Nicaragua (10), and El Salvador (1). The Caribbean region also exhibits limited scientific production, with countries like Barbados (13 articles), Antigua (54), and Dominica (1) showing modest contributions. These results highlight significant disparities in research capacity across LAC, likely driven by differences in research funding, institutional development, and national energy policies.

Beyond the LAC region, international contributions to renewable energy research are also noteworthy. The United States (647 articles), Spain (630), and the United Kingdom (210) have been key contributors, suggesting strong collaborative ties between LAC institutions and leading international research centers. Other notable contributors include Germany (185), China (175), France (163), Italy (142), India (134), and Portugal (126), reflecting a globalized research landscape where energy sustainability is a shared priority.

Table 3. Top 30 scientific journals for research on renewable energies. Hirsch index = h-index, TC = total citations, NP = number of publications, and PY_start = year of publication start. The table is organized in descending order by NP.

Source	TC	NP	h-Index	PY_start
hEnergies	7238	745	35	2010
Renewable Energy	19,492	638	70	1994
Solar Energy	14,416	442	56	1994
Renewable & Sustainable Energy Reviews	17,108	312	65	2000
Energy	8896	280	50	1994
IEEE Latin America Transactions	1629	275	18	2007
Sustainability	2284	241	23	2014
Renewable Energy and Power Quality Journal	408	208	9	2005
Journal of Cleaner Production	5922	194	44	2007
International Journal of Hydrogen Energy	4413	185	35	1998
Electric Power Systems Research	2726	181	27	1997
Applied Energy	6972	163	46	2003
Energy Policy	5325	157	42	1994
Energy Conversion and Management	5171	151	42	2000
IEEE Access	1994	144	23	2016
Solar Energy Materials and Solar Cells	6041	137	45	1994
Applied Sciences-Basel	1122	119	17	2018
Science of the Total Environment	3441	110	37	2013
Advances in Space Research	1056	107	19	1999
Biomass & Bioenergy	2627	98	30	2003
International Journal of Electrical Power & Energy Systems	1542	96	24	2008
Journal of Atmospheric and Solar-Terrestrial Physics	1509	92	23	1997
IEEE Transactions on Industrial Electronics	12,814	87	59	2001
Sustainable Energy Technologies and Assessments	1454	87	21	2016
Bioenergy Research	1341	87	19	2013
Applied Thermal Engineering	1593	83	23	1998
Bioresource Technology	5979	76	43	2007
IEEE Transactions on Power Systems	5353	75	43	1994
Industrial Crops and Products	1990	73	25	2011
Brazilian Archives of Biology and Technology	110	71	5	2000

presents the 30 most relevant scientific journals for renewable energy research, selected from a total of 3193 sources. The list ranks journals based on their number of publications (NP) while also considering total citations (TC), h-index (Hirsch index), and year of publication start (PY_start). These indicators provide insights into the leading publication venues and their scientific influence.

At the top of the list, Energies ranks first with 745 publications, 7238 citations, and an h-index of 35, since its launch in 2010. Renewable Energy, which started in 1994, ranks second in publication volume (638 articles) but stands out with 19,492 citations and the highest h-index (70), indicating its substantial impact. Other key journals include Solar Energy (h-index 56, 14,416 citations) and Renewable & Sustainable Energy Reviews (h-index 65, 17,108 citations), demonstrating a strong balance between publication frequency and academic recognition.

In addition to global journals, regional sources such as IEEE Latin America Transactions (275 articles, h-index 18) highlight the role of Latin American journals in disseminating local research. However, their citation impact remains lower compared to global publications. Similarly, Renewable Energy and Power Quality Journal (208 articles, h-index 9) has made contributions since 2005 but shows limited citation reach.

Among interdisciplinary journals, the Journal of Cleaner Production has gained relevance in sustainability research, with 194 publications, 5922 citations, and an h-index of 44. Other journals such as Sustainability (241 articles, 2284 citations) and Science of the Total Environment (110 articles, 3441 citations) reflect the broadening scope of renewable energy research into socio-environmental aspects.

Technical journals such as Applied Energy (163 articles, h-index 46) and Energy Policy (157 articles, h-index 42) focus on energy management and policy. Meanwhile, IEEE Transactions on Industrial Electronics (87 articles, 12,814 citations, h-index 59) demonstrates a high impact despite lower publication volume, indicating a strong influence in energy systems and smart grid research.

Finally, local journals such as Brazilian Archives of Biology and Technology (71 articles, h-index 5) contribute to regional knowledge dissemination but face challenges in achieving global visibility. This contrast between international and regional journals highlights the need for stronger institutional support to enhance the visibility and impact of Latin American research.

3.2. Latent Dirichlet Allocation (LDA)

The application of LDA to the corpus of 18,780 scientific publications revealed 18 dominant topics (k = 18) as the optimal configuration, determined based on the highest coherence score. This suggests that renewable energy research in Latin America and the Caribbean (LAC) can be effectively categorized into 18 thematic areas, encompassing a diverse range of subjects from technological advancements in energy systems to policy frameworks and environmental impact assessments. Consequently, the solution with k = 18 provided the optimal balance. It offered sufficient thematic resolution to distinguish between critical domains (e.g., separating “Materials Research for Solar Cells” from “Microgrid Optimization”), while ensuring that each topic remained coherent and interpretable. This level of granularity is particularly valuable for policy interpretation, as it allows the identification of both technology-specific advances and systemic issues, including governance, financing, and regional disparities. Finally, the suitability of the 18-topic model was validated during the expert manual labeling process, in which each topic was assigned a clear semantic label based on its most frequent terms and representative documents. This confirmed that the topics were both intellectually coherent and non-redundant. Thus, the selection of k = 18 reflects not only statistical optimization but also scholarly interpretability and practical relevance, ensuring that the model provides meaningful insights for understanding the evolution of renewable energy research in LAC.

Table 4. Identified topics in renewable energy research in Latin America and the Caribbean using LDA.

Topic	Top_Terms	N	Label
t_1	process, treatment, product, organ, concentr, remov, degrad, soil, rate, produc, reactor, bioga, wast, condit, digest	653	Waste Treatment and Conversion Processes
t_2	project, analysi, econom, methodologi, decis, plan, risk, assess, program, market, base, select, evalu, paper, consid	274	Economic Evaluation and Energy Project Management
t_3	temperatur, thermal, heat, effici, degre, perform, concentr, condit, air, water, solar, design, experiment, collector, rate	1070	Solar Technologies and Photovoltaic Energy
t_4	activ, increas, mitochondri, effect, cell, function, level, induc, decreas, mechan, metabol, respons, protein, stress, complex	791	Energy Storage and Bioenergy
t_5	control, power, propos, voltag, current, grid, convert, paper, oper, dc, connect, base, gener, strategi, design	1686	Control and Optimization of Energy Systems
t_6	water, river, speci, reservoir, plant, hydroelectr, dam, hydropow, region, brazil, fish, impact, basin, construct, flow	1788	Environmental Impact of Hydropower and Aquatic Ecosystems
t_7	model, method, base, data, predict, propos, time, estim, approach, perform, paramet, appli, algorithm, optim, compar	1208	Modeling and Prediction Using Artificial Intelligence in Renewable Energy
t_8	environment, impact, emiss, sustain, climat, fuel, assess, scenario, chang, countri, potenti, global, sector, reduc, ga	1137	Climate Change and Decarbonization Policies
t_9	data, observ, region, measur, period, time, averag, variabl, variat, valu, correl, activ, analysi, estim, station	832	Interaction Between Solar Activity and Earth’s Electromagnetic Field
t_10	wind, turbin, wind_turbin, speed, power, farm, wind_energi, wind_power, wind_speed, wind_farm, blade, design, simul, structur, dynam	1012	Wind Turbines and Optimization of Wind Energy Systems
t_11	energi, electr, gener, renew, sourc, renew_energi, consumpt, demand, energi_sourc, potenti, technologi, altern, develop, suppli, power	1435	Development and Overview of Renewable Energy in Latin America
t_12	product, biomass, produc, bioenergi, potenti, sugarcan, content, evalu, yield, crop, plant, residu, process, soil, oil	1694	Biomass Production and Utilization for Energy
t_13	cell, effici, devic, solar_cell, electron, convers, photovolta, solar, light, base, perform, densiti, charg, organ, layer	1424	Materials Research for Solar Cells and Optoelectronics
t_14	photovolta, pv, system, modul, perform, condit, panel, photovolta_pv, instal, arrai, monitor, measur, detect, pv_system, photovolta_system	445	Diagnostics and Maintenance of Photovoltaic Modules
t_15	film, structur, materi, properti, surfac, deposit, optic, tio, character, oxid, layer, applic, sampl, rai, spectroscopi	1217	Nanostructures and Thin Films for Solar Energy
t_16	solar, solar_energi, radiat, irradi, hydrogen, solar_radiat, solar_irradi, energi, dai, sun, solar_power, surfac, averag, direct, incid	254	Evaluation and Modeling of Solar Radiation
t_17	power, cost, system, oper, gener, optim, plant, storag, distribut, integr, hybrid, batteri, solut, grid, propos	976	Optimization and Management of Microgrids
t_18	develop, review, identifi, discuss, studi, commun, provid, challeng, aim, process, main, literatur, technologi, local, practic	884	Sustainability and Governance in Renewable Energy Projects

presents the 18 identified topics, along with their most representative terms, number of associated documents (N), and thematic labels. These topics reflect the multidisciplinary nature of renewable energy research in LAC, integrating engineering, environmental sciences, policy studies, and energy economics. Notably, the research landscape is predominantly shaped by technological advancements, with certain topics attracting significantly more scholarly attention than others.

A closer analysis of the distribution of topics highlights the key focus areas within renewable energy research in LAC. Among the most prominent research topics, the environmental impact of hydropower and aquatic ecosystems (t_6, N = 1788) emerges as the most extensively studied area, emphasizing the ecological consequences of hydroelectric projects and their impact on riverine biodiversity. This is followed closely by biomass production and utilization for energy (t_12, N = 1694), underlining the region’s growing interest in bioenergy solutions. Additionally, the control and optimization of energy systems (t_5, N = 1686) and materials research for solar cells and optoelectronics (t_13, N = 1424) signal a strong technological emphasis on energy management and photovoltaic advancements.

Conversely, topics related to policy and governance occupy a relatively smaller share of the research landscape. For instance, sustainability and governance in renewable energy projects (t_18, N = 884) and economic evaluation and energy project management (t_2, N = 274), though essential for the successful implementation of energy transitions, are underrepresented compared to engineering-focused topics. This suggests a research gap in addressing financial mechanisms, regulatory frameworks, and socio-political challenges associated with renewable energy deployment in LAC.

Furthermore, the presence of topics such as interaction between solar activity and Earth’s electromagnetic field (t_9, N = 832) and evaluation and modeling of solar radiation (t_16, N = 254) indicates that a segment of researchers is engaged in fundamental studies on solar energy variability, though these areas remain less dominant compared to applied renewable energy research.

The LDA results reveal four overarching thematic clusters within the renewable energy research landscape in LAC:

• Technological Innovations in Renewable Energy Systems: Topics such as solar technologies (t_3), wind energy (t_10), bioenergy (t_12), and microgrid optimization (t_17) highlight the region’s focus on developing and improving renewable energy technologies. The strong emphasis on materials science (t_13, t_15) further suggests an interest in enhancing photovoltaic performance and energy conversion efficiency.
• Environmental and Ecological Impact of Renewable Energy: The dominance of hydropower impact studies (t_6) underscores the region’s reliance on hydroelectric energy and the growing concern over its ecological consequences. Similarly, research on climate change mitigation (t_8) and solar radiation modeling (t_16) reflects efforts to assess renewable energy’s role in decarbonization.
• Energy Policy, Economics, and Governance: Despite the dominance of technological research, the presence of policy-related topics (t_18, t_2) highlights an emerging recognition of the socioeconomic dimensions of energy transition. However, their lower document count suggests that policy frameworks, financial mechanisms, and governance strategies remain underexplored compared to engineering solutions.
• Artificial Intelligence and Computational Modeling in Renewable Energy: The integration of AI in energy systems (t_7) and grid optimization strategies (t_5, t_17) signals a transition towards data-driven decision-making and automation in energy management.

Thereby, the LDA-based topic modeling provides a structured mapping of renewable energy research in LAC, revealing key focus areas, emerging trends, and research gaps. The dominance of technological advancements suggests that research efforts are primarily driven by engineering and applied sciences, while policy-oriented studies and economic analyses remain secondary. This highlights the need for greater interdisciplinary collaboration to ensure a holistic approach to the region’s energy transition.

3.3. Topics Trends in Renewable Energy Research

The temporal analysis of research trends in renewable energies reveals distinct patterns in scientific knowledge evolution across different thematic areas. Some topics exhibit a positive trajectory, indicating an increase in interest and academic output, while others display declining trends, suggesting a shift in research priorities. Additionally, several topics remain relatively stable or exhibit fluctuations without a clear directional trend. Figure 3 provides a visual summary of these trends, where topics experiencing sustained growth are marked in red, declining topics in blue, and fluctuating or stable topics in black.

3.3.1. Growing Research Topics in Renewable Energy

The analysis identified six research areas with a sustained upward trend, reflecting emerging priorities in the field of renewable energy. One of the most prominent growing fields is modeling and prediction using artificial intelligence in renewable energy (t_7). The integration of machine learning techniques and AI-driven predictive models has gained increasing relevance in recent years. These methodologies are optimizing energy management by improving demand forecasting accuracy and enhancing the efficiency of energy generation systems. Moreover, the adoption of these tools has facilitated the automation of power grids and the optimization of energy distribution, contributing to the stability and resilience of smart grids.

Simultaneously, the formulation and implementation of decarbonization policies and climate change mitigation strategies (t_8) has experienced significant growth. Research in this area focuses on carbon emission reduction, the design of mitigation strategies, and the development of sustainable energy policies. This trend aligns with international commitments to achieve net-zero emissions and strengthen the global energy transition.

Another rapidly expanding area is waste treatment and energy conversion processes (t_1). The increasing volume of research in this field reflects the growing interest in energy recovery solutions, such as biogas production and waste-to-energy conversion. These strategies not only support sustainable energy generation but are also embedded within the principles of the circular economy.

Additionally, the study of biomass production and utilization for energy (t_12) has seen remarkable expansion, particularly in research on agricultural waste processing and bioenergy production from crops such as sugarcane. This growth highlights Latin America and the Caribbean’s commitment to developing alternative fuels and efficiently utilizing renewable natural resources.

Research on the development and implementation of renewable energy in Latin America (t_11) has also gained increasing academic importance. This research line focuses on assessing the region’s energy potential and analyzing its role in the global energy transition. It also emphasizes the importance of regional case studies to understand the challenges and opportunities related to renewable energy adoption in diverse socioeconomic contexts.

Finally, interest in the optimization and management of microgrids (t_17) has grown considerably. The decentralization of energy systems and microgrid optimization have become central to discussions on energy resilience and storage solutions. The increasing integration of hybrid and storage solutions in decentralized grids demonstrates the need for more efficient and flexible strategies for energy management in isolated communities and urban settings.

The growth of these research areas suggests an increasing focus on alternative energy sources, efficiency improvements in decentralized energy systems, and the strengthening of sustainable technological strategies for renewable energy generation and management in Latin America and the Caribbean.

3.3.2. Declining Research Topics in Renewable Energy

In contrast to the growing research areas, the analysis identified five topics that have exhibited a downward trend in publication volume over recent years. This decline may indicate a maturing field, a shift in research priorities, or a reduction in funding allocation.

One of the areas experiencing a decrease in academic interest is economic evaluation and energy project management (t_2). The declining focus on financial and economic analyses of energy projects suggests that existing models may have reached a saturation point, prompting a shift toward more technical and policy-driven approaches to renewable energy planning.

Similarly, solar technologies and photovoltaic energy (t_3), despite being a cornerstone of renewable energy, have shown a reduction in publication output. While photovoltaics remain a key energy source, the decline in fundamental research in this area indicates a transition from theoretical advancements to applied research and large-scale commercial implementation.

Another topic that has seen a diminishing research focus is the interaction between solar activity and Earth’s electromagnetic field (t_9). Studies on solar variability and its impact on energy systems have become less frequent, likely due to the limited practical applications of these investigations compared to other energy modeling approaches.

A notable decrease has also been observed in research on nanostructures and thin films for solar energy (t_15). The declining interest in nano-optimized solar materials suggests a shift in focus towards large-scale photovoltaic technologies and next-generation materials with broader industry applications.

Lastly, the field of evaluation and modeling of solar radiation (t_16), once crucial for optimizing solar energy efficiency, has experienced a reduction in academic contributions. Advances in computational techniques and the availability of standardized datasets may have contributed to this decline, as improved predictive models have lessened the need for further foundational studies.

These declining research trajectories reflect a realignment of scientific priorities, where attention is shifting from fundamental investigations toward applied and system-level advancements in renewable energy. This transition highlights the growing emphasis on implementation, scalability, and integration of renewable energy solutions into practical energy systems.

3.3.3. Stable and Fluctuating Research Topics

Several remaining topics exhibit periodic fluctuations without a clear increasing or decreasing trend. These include research on hydropower impact assessments, control and optimization of energy systems, and diagnostics for photovoltaic technologies. The intermittent variations in publication volume may reflect external factors such as funding availability, technological advancements, or policy shifts. The continued presence of these topics suggests that they remain of scientific interest but are subject to variability in research output rather than sustained growth.

Therefore, the temporal evolution of research topics in renewable energy demonstrates a dynamic landscape, where certain areas are expanding rapidly, while others are experiencing decline or stabilization. The rise of AI-driven modeling, bioenergy, and microgrid optimization signals an increasing emphasis on digitalization, sustainability, and decentralized energy solutions. Conversely, the decreasing focus on traditional photovoltaic materials, solar activity studies, and energy economics suggests that these areas may have reached technological maturity or face shifting funding priorities.

Future research efforts should consider bridging the gap between technological advancements and policy implementation, ensuring that emerging innovations translate into real-world energy transitions in the LAC region.

3.4. Visualization of the Intertopic Distance Map

The Intertopic Distance Map (Figure 4), generated via Multidimensional Scaling (MDS), offers a conceptual framework for interpreting the intellectual structure of renewable energy research in LAC. The horizontal axis (PC1) can be understood as a “Technology-to-Implementation” spectrum, separating topics grounded in fundamental materials science and technological innovation (left) from those addressing policy, economics, and governance (right). The vertical axis (PC2) differentiates methodological paradigms, with computational and optimization studies concentrated in the upper region, and ecological and sustainability-oriented research in the lower region. The spatial distribution of topic clusters provides meaningful analytical insights. On the far-left of PC1, a dense cluster of technology-driven topics—such as materials for solar cells (t_13), nanostructures (t_15), and solar technologies (t_3)—represents a research paradigm centered on fundamental innovation and efficiency gains. At the opposite end of the axis, the grouping of renewable energy development overviews (t_11) and economic evaluation (t_2) reflects a paradigm focused on implementation feasibility, financing, and policy frameworks. The significant separation between these two poles visually demonstrates the enduring science–policy gap: technological discovery and socio-economic integration remain largely disconnected research domains. The vertical axis reinforces this duality. The upper region is dominated by computationally intensive topics such as AI-driven prediction (t_7), control and optimization of systems (t_5), and wind turbine optimization (t_10), highlighting a methodological emphasis on digitalization and performance optimization. Conversely, the lower region groups topics like hydropower impacts on ecosystems (t_6) and biomass utilization (t_12), representing research concerned with sustainability and environmental consequences. This divergence suggests that LAC’s research landscape tends to treat efficiency and ecological impacts as separate streams, with relatively few integrative approaches bridging these two critical dimensions.

Thematic Organization of Research Topics

The horizontal axis (PC1) represents a thematic progression from technological research on energy conversion and materials development (left side) to economic feasibility analyses and policy studies (right side). The vertical axis (PC2) differentiates computational and system optimization studies (upper region) from sustainability and environmental impact assessments (lower region). Several important aspects of PC1 and PC2 are the following:

• Technological and materials research for renewable energy (left side of PC1)

A distinct cluster on the left side of PC1 is dedicated to research on solar technologies and photovoltaic materials. This group encompasses key areas such as materials research for solar cells and optoelectronics (t_13), nanostructures and thin films for solar energy (t_15), solar technologies and photovoltaic energy (t_3), and evaluation and modeling of solar radiation (t_16). The concentration of these topics suggests that solar energy research in Latin America and the Caribbean is strongly oriented toward enhancing photovoltaic conversion efficiency through materials engineering and solar radiation assessment. Notably, the close relationship between advancements in nanotechnology (t_15) and photovoltaic optimization (t_13) highlights an ongoing effort to improve solar energy performance using innovative materials. This emphasis on nanostructured materials and thin-film technologies underscores the scientific community’s commitment to developing more efficient and sustainable solar energy solutions.

• Economic and policy studies in renewable energy (right side of PC1)

At the opposite end of PC1, a distinct cluster emerges, encompassing topics related to the implementation of renewable energy, economic feasibility, and policy considerations. This group includes the development and overview of renewable energy in Latin America (t_11) alongside economic evaluation and energy project management (t_2). The proximity of these topics suggests that the success of renewable energy transitions in Latin America and the Caribbean is strongly influenced by economic and regulatory factors. While technological advancements serve as key drivers of innovation, their large-scale application and long-term viability depend on well-structured financial frameworks, strategic investments, and robust policy development. The strong interconnection between these areas underscores the need for integrated approaches that effectively align economic planning with the deployment of renewable energy technologies, ensuring both feasibility and sustainability in the region’s energy transition.

• Computational methods and energy system optimization (upper region of PC2)

The upper section of PC2 is characterized by a cluster of topics centered on computational modeling and optimization, with a strong emphasis on improving energy management and forecasting. This group includes control and optimization of energy systems (t_5), modeling and prediction using artificial intelligence in renewable energy (t_7), wind turbines and optimization of wind energy systems (t_10), and diagnostics and maintenance of photovoltaic modules (t_14). The close relationship between energy system optimization (t_5) and AI-driven modeling (t_7) highlights the increasing application of artificial intelligence in optimizing energy distribution, efficiency, and system reliability. Similarly, the strong connection between wind energy optimization (t_10) and AI-based prediction models (t_7) suggests that advanced computational techniques are being applied across multiple renewable energy sources, particularly wind and solar power systems. These interrelations reinforce the growing shift toward data-driven energy management, where machine learning and predictive analytics play a crucial role in enhancing system performance, stability, and long-term sustainability.

• Environmental sustainability and ecological impact (lower region of PC2)

At the lower extreme of PC2, a cluster of topics is centered on environmental assessments and sustainability, emphasizing the ecological and bioenergy aspects of renewable energy systems. This group includes the environmental impact of hydropower and aquatic ecosystems (t_6), biomass production and utilization for energy (t_12), and energy storage and bioenergy (t_4). The positioning of hydropower’s environmental impact (t_6) at the lowest point of PC2 suggests that its focus is distinct from technological and computational studies, as it primarily addresses the ecological consequences of hydropower infrastructure on biodiversity, aquatic ecosystems, and water resource management.

Meanwhile, the close relationship between biomass utilization (t_12) and energy storage solutions (t_4) highlights the strong interconnection between bioenergy research and advancements in energy storage and conversion efficiency. This link underscores the growing interest in optimizing biomass-derived energy for long-term sustainability, reinforcing the role of bioenergy as both a renewable resource and a complementary solution to energy storage challenges.

• Intermediate topics and cross-disciplinary connections

Some topics occupy an intermediate position in the MDS plot, serving as bridges between different research domains:

• Interaction Between Solar Activity and Earth’s Electromagnetic Field (t_9) is positioned between solar radiation studies and broader geophysical processes, indicating its dual relevance to both renewable energy modeling and atmospheric sciences.
• Evaluation and Modeling of Solar Radiation (t_16) connects solar technology research with energy system integration, reflecting the critical role of radiation assessment in optimizing solar energy output.

The Intertopic Distance Map provided valuable insights into the structural relationships among renewable energy research themes in LAC. It reveals three major clusters:

• Technological advancements in solar and nanomaterials research (left side of PC1)
• Economic and policy-driven energy studies (right side of PC1)
• Computational modeling and AI-driven optimization (upper PC2) vs. environmental impact and sustainability assessments (lower PC2)

This spatial distribution highlights the interdisciplinary nature of renewable energy research, where technological, economic, and environmental factors are interconnected. Moving forward, enhanced collaboration between material scientists, engineers, policymakers, and environmental researchers will be essential to bridge knowledge gaps and support a sustainable energy transition in LAC.

3.5. HJ-Biplot Analysis

The HJ-Biplot (Figure 5) provides a two-dimensional representation of the evolution of research topics in renewable energy in LAC. The first principal component (PC1) explains 42.19% of the variability, while the second component (PC2) captures an additional 18.52%, suggesting that this plane effectively reflects a significant portion of research evolution. By analyzing the positioning of topics and years, we can identify shifts in research priorities over time. Hence, the HJ-Biplot provides a temporal lens through which the evolution of research priorities can be understood. Rather than serving as a static snapshot, the visualization traces a narrative of maturation, showing how the field has shifted from a technology-centric foundation toward a more systemic and policy-aware research agenda.

3.5.1. Early Research Focus (1994–2003): Foundations in Science and Technology

At the far left of the biplot, corresponding to the period from 1994 to 2003, research was primarily focused on solar energy and material science. This early stage is characterized by a cluster of topics, including nanostructures and thin films for solar energy (t_15), evaluation and modeling of solar radiation (t_16), solar technologies and photovoltaic energy (t_3), and interaction between solar activity and Earth’s electromagnetic field (t_9). The emphasis during this phase was on fundamental research, particularly in the development of advanced materials for photovoltaic applications, the characterization of solar radiation, and the study of solar activity’s impact on Earth’s energy systems. These investigations provided the scientific foundation for subsequent technological advancements in solar power generation. The concentration of these topics in the leftmost region of PC1 suggests that early research was highly specialized and technologically focused, with limited interdisciplinary integration.

It is then possible to affirm that the first stage of renewable energy research in LAC was dominated by topics such as solar energy materials (t_15), radiation modeling (t_16), and early photovoltaic technologies (t_3). This phase reflects an essential effort to establish baseline scientific and technological knowledge, focusing on material properties, resource characterization, and the feasibility of solar-based systems.

3.5.2. Technological Expansion and Diversification (2007–2014)

As research progressed between 2007 and 2014, a noticeable shift in focus emerged, particularly in the upper region of the biplot, where topics related to renewable energy implementation and optimization became more prominent. This period saw increased emphasis on biomass production and utilization for energy (t_12) and diagnostics and maintenance of photovoltaic modules (t_14). The transition from theoretical exploration to applied research became evident, with efforts directed toward improving the efficiency of biomass-based energy production and enhancing the reliability of photovoltaic systems.

The growing attention to biomass energy suggests an expansion of the research scope beyond solar energy, acknowledging biomass as a viable and increasingly relevant renewable resource in Latin America and the Caribbean. The positioning of these topics along PC1 indicates that this phase served as a bridge between fundamental materials research and large-scale energy deployment, marking a crucial period of technological optimization and the development of maintenance strategies essential for the long-term sustainability of renewable energy systems.

This means that the subsequent period was marked by a transition to applied research, with increasing emphasis on operational reliability and diversification of renewable sources. Topics such as biomass utilization (t_12) and diagnostics and maintenance of PV modules (t_14) gained prominence, reflecting a pivot toward practical deployment and system optimization. This period signaled the region’s growing capacity to move from theoretical exploration to the applied engineering challenges of scaling renewable technologies.

3.5.3. Recent Research Trends (2016–2024): Policy, Decentralization, and Energy Transition

In the most recent phase, coinciding with global climate milestones such as the Paris Agreement and the SDGs, the focus has expanded to include climate change and decarbonization policies (t_8), regional overviews of renewable energy (t_11), and microgrid optimization (t_17). This trajectory illustrates a paradigm shift: research is no longer confined to technological innovation but increasingly situates renewable energy within broader socio-economic and governance frameworks. The prominence of decentralized systems, resilience, and policy analysis indicates that the research community now recognizes the energy transition as a multi-dimensional challenge requiring the integration of technology, governance, and social equity.

Taken together, the MDS and HJ-Biplot analyses reveal that renewable energy research in LAC has progressed along a path from scientific discovery to applied engineering and, more recently, toward systemic and policy-oriented integration. These visualizations thus provide more than descriptive overviews; they highlight structural divides, historical shifts, and emerging synergies that collectively define the region’s evolving research agenda.

Therefore, beyond their descriptive value, the MDS and HJ-Biplot analyses provide deeper analytical insights into the intellectual structure and temporal evolution of renewable energy research in LAC. The MDS intertopic distance map revealed a clear thematic polarization, with one axis differentiating technological and computational advances from policy-oriented and sustainability-focused studies, and the other axis distinguishing optimization-driven approaches from environmental impact assessments. This spatial configuration suggests that research in the region has evolved along two fundamental tensions: innovation versus governance, and efficiency versus sustainability. Similarly, the HJ-Biplot allowed us to uncover not only chronological shifts in research priorities but also the specific variables driving these transitions. The trajectories observed demonstrate how early research centered on solar materials and radiation gradually gave way to applied biomass and photovoltaic maintenance studies, and more recently to systemic issues such as climate policy, decentralization, and energy transition. These analytical patterns highlight that the knowledge structure of renewable energy research in LAC is not static but reflects a progressive integration of technological, environmental, and socio-political dimensions, an insight that would not have emerged from descriptive statistics alone.

3.5.4. Key Insights from the HJ-Biplot

• Early-stage research (1994–2003) was highly specialized, focusing on solar technologies, material science, and radiation modeling, laying the scientific foundation for future advancements.
• The period from 2007–2014 marked a transition toward applied research, emphasizing biomass energy production and photovoltaic maintenance, suggesting a shift from theory to practical implementation.
• Recent research (2016–2024) reflects an interdisciplinary approach, integrating climate policies, energy transition strategies, and decentralized grid optimization, indicating a broad and systemic focus.

3.6. Collaboration Networks in Renewable Energy Research in LAC

Scientific collaboration plays a critical role in advancing research and innovation, particularly in multidisciplinary fields such as renewable energy. Understanding the structure of collaboration networks provides valuable insights into the dynamics of knowledge production, the level of integration between research institutions, and the international linkages that shape research output. This section presents the results of a bibliometric analysis of collaboration networks in renewable energy research in LAC, highlighting key countries.

3.6.1. Structure of the Collaboration Network

The network analysis of scientific publications in renewable energy across LAC reveals a highly interconnected structure, where collaboration occurs at national, regional, and international levels. The analysis identified four major clusters of collaboration, each exhibiting distinct characteristics in terms of centrality, influence, and connectivity:

• Cluster 1: Global Research Hubs

This cluster includes highly central and influential countries, which act as key nodes in the research network. The most dominant actors are:

• Brazil (Betweenness: 451,466)
• United States (379,247)
• United Kingdom (193,970)
• Germany (123,628)

Brazil stands out as the leading country in renewable energy research in LAC (

Table A1. Collaborative networks for research on renewable energy.

Node	Cluster	Betweenness	Closeness	PageRank
Brazil	1	451,466	0.009	0.091
USA	1	379,247	0.009	0.064
United Kingdom	1	193,970	0.009	0.042
Germany	1	123,628	0.009	0.036
Italy	1	91,212	0.009	0.025
Portugal	1	5036	0.008	0.012
Canada	1	19,093	0.008	0.021
Netherlands	1	41,206	0.008	0.02
Australia	1	33,615	0.008	0.02
Costa Rica	1	2885	0.007	0.006
Sweden	1	18,824	0.008	0.016
Switzerland	1	38,019	0.008	0.017
Japan	1	10,436	0.008	0.011
Denmark	1	9254	0.008	0.012
Austria	1	11,103	0.008	0.013
Poland	1	8597	0.008	0.009
Norway	1	5849	0.008	0.01
Finland	1	7517	0.007	0.009
Czech Republic	1	2234	0.007	0.005
Ireland	1	3108	0.007	0.006
Panama	1	674	0.006	0.003
Paraguay	1	14	0.006	0.002
Turkey	1	1784	0.007	0.005
Greece	1	2877	0.007	0.005
Israel	1	771	0.007	0.004
Romania	1	1787	0.007	0.005
Ukraine	1	16	0.006	0.003
New Zealand	1	1066	0.007	0.005
Barbados	1	63	0.006	0.002
Croatia	1	549	0.006	0.004
Hungary	1	2665	0.007	0.004
Nicaragua	1	4	0.005	0.002
Mozambique	1	0	0.005	0.002
Serbia	1	889	0.006	0.003
Slovenia	1	295	0.006	0.003
Estonia	1	42	0.006	0.003
Slovakia	1	9	0.006	0.002
Bulgaria	1	14	0.006	0.002
Kenya	1	104	0.006	0.003
Nepal	1	83	0.006	0.002
Lithuania	1	251	0.006	0.003
Sri Lanka	1	0	0.005	0.002
Belize	1	0	0.005	0.002
Mexico	2	212,024	0.009	0.045
Chile	2	201,932	0.009	0.045
Colombia	2	41,623	0.008	0.026
Spain	2	181,373	0.009	0.053
Argentina	2	4522	0.007	0.013
Ecuador	2	16,412	0.007	0.015
France	2	102,455	0.009	0.026
Cuba	2	1233	0.007	0.007
Uruguay	2	789	0.007	0.005
Venezuela	2	806	0.006	0.004
Belgium	2	18,182	0.007	0.009
Morocco	2	511	0.006	0.004
Algeria	2	3516	0.006	0.004
Bolivia	2	274	0.006	0.004
Jamaica	2	768	0.006	0.003
Suriname	2	21	0.005	0.002
Luxembourg	2	21	0.006	0.002
Ghana	2	0	0.005	0.002
Togo	2	0	0.005	0.002
China	3	90,804	0.009	0.028
India	3	78,998	0.008	0.022
Peru	3	26,476	0.008	0.011
Saudi Arabia	3	33,985	0.008	0.017
Pakistan	3	5864	0.007	0.01
Egypt	3	7495	0.007	0.008
Iran	3	6934	0.008	0.008
Malaysia	3	5364	0.007	0.007
Korea	3	3604	0.007	0.009
South Africa	3	9371	0.007	0.007
Singapore	3	1533	0.007	0.006
Russia	3	6349	0.008	0.007
Iraq	3	6090	0.007	0.005
U Arab Emirates	3	5214	0.007	0.006
Vietnam	3	1339	0.007	0.004
Thailand	3	1754	0.007	0.005
Tunisia	3	200	0.006	0.004
Indonesia	3	3692	0.007	0.005
Ethiopia	3	4461	0.007	0.004
Nigeria	3	725	0.006	0.003
Jordan	3	719	0.006	0.003
Bangladesh	3	15	0.006	0.002
Cameroon	3	66	0.006	0.003
Qatar	3	578	0.006	0.003
Azerbaijan	3	323	0.006	0.003
Lebanon	3	165	0.006	0.003
Philippines	3	159	0.006	0.003
Uzbekistan	3	184	0.006	0.002
Kuwait	3	145	0.006	0.003
Cyprus	3	30	0.006	0.003
Uganda	3	122	0.006	0.002
Oman	3	2	0.005	0.002
Guatemala	4	124	0.006	0.002
Honduras	4	397	0.006	0.003
El Salvador	4	0	0.005	0.002

, Appendix A), serving as a bridge between regional and global research efforts. The strong presence of the United States and European nations in this cluster reflects their role in funding, joint research projects, and technological collaboration with LAC institutions. These countries facilitate knowledge transfer and provide access to advanced research infrastructure.

• Cluster 2: Regional Leaders

This group consists of the most prolific Latin American countries in renewable energy research:

• Mexico (212,024)
• Chile (201,932)
• Colombia (41,623)
• Spain (181,373)

The significant presence of Spain in this cluster suggests historical and academic ties with Latin America, fostering joint research initiatives and institutional collaborations. Other regional contributors include Argentina and Ecuador, though with lower influence compared to leading nations.

• Cluster 3: Emerging Collaborators

This cluster consists of countries that are increasing their participation in renewable energy research, including:

• China (90,804)
• India (78,998)
• Peru (26,476)

The growing interest of China and India in LAC’s renewable energy sector aligns with their global strategies for expanding technological influence and securing renewable energy investments. Additionally, countries such as Saudi Arabia and South Africa appear in this cluster, reflecting their sporadic but notable collaborations with LAC institutions.

• Cluster 4: Low-Connectivity Countries

This final cluster includes countries with limited participation in renewable energy research and low connectivity within the network, such as:

• Guatemala
• Honduras
• El Salvador

The low centrality values of these nations indicate a lack of integration into major regional or international research collaborations, likely due to limited funding, institutional capacity, or national policy prioritization. The findings highlight persistent disparities in research collaboration, emphasizing the need for stronger regional integration to foster a more cohesive and impactful research ecosystem.

3.6.2. Key Network Metrics and Country Influence

To quantify the influence and connectivity of different actors in the network, centrality measures such as PageRank, Closeness, and Betweenness were computed. PageRank values indicate the most central and influential countries in the network, with:

• Brazil (0.091), Mexico (0.045), and Chile (0.045) ranking highest.
• Spain (0.053), despite being outside LAC, holds a strong position due to its extensive collaborations with Latin American institutions.

Betweenness centrality highlights Brazil, Mexico, and the United States as essential intermediaries facilitating knowledge exchange between different parts of the network. The presence of high-income countries (USA, Germany, UK) as dominant actors underscores the reliance of LAC researchers on collaborations with well-funded international institutions. These metrics reveal that while some LAC nations play a leading role in research collaboration, there is still a heavy dependence on global research hubs for funding and expertise.

4. Discussion

This study employs a robust methodological framework, integrating machine learning techniques to systematically assess the evolution of scientific production on renewable energy in LAC. The use of Latent Dirichlet Allocation (LDA) topic modeling allowed for the identification of emerging research themes, offering a data-driven approach to understanding the intellectual structure of the field. Additionally, network analysis provided insights into collaboration patterns and the degree of regional integration in scientific research. These methodological approaches enabled a more comprehensive evaluation of research trends, moving beyond traditional literature reviews by leveraging large-scale data analysis to uncover latent patterns and thematic shifts in renewable energy studies.

The methodological approach adopted in this study is particularly valuable for capturing complex and dynamic research trends. LDA topic modeling enables the identification of underlying thematic structures within large datasets, providing an objective perspective on the evolution of scientific discourse. Unlike conventional literature reviews, which may be limited by subjective selection biases, this technique facilitates a more systematic and reproducible means of detecting shifts in research priorities over time. Similarly, network analysis offers a quantitative assessment of collaboration structures, highlighting disparities in scientific integration across countries. These methodological tools not only enhance our understanding of renewable energy research in LAC but also provide a framework that can be applied to other emerging fields requiring systematic trend analysis. In this manner, this study provides a comprehensive examination of the evolution of scientific production on renewable energy in LAC, identifying key trends that define the region’s renewable energy research landscape. The findings reveal a field characterized by sustained growth, clear thematic evolution, and persistent structural imbalances. These results are best interpreted in dialogue with the broader literature on scientific production and the global energy transition. Overall, the expansion of research output in LAC is consistent with international trajectories, aligning with the Sustainable Development Goals (SDG 7 on Affordable and Clean Energy, and SDG 13 on Climate Action), where innovations in renewable energy are increasingly recognized as fundamental enablers of sustainable development [45].

Although renewable energy research in LAC has grown substantially, the effective translation of scientific outputs into actionable policies and practical implementations remains limited. Our findings reveal a misalignment between the concentration of research on technological innovations and the relatively weaker focus on the socio-economic, governance, and financial frameworks needed for a sustainable energy transition. Bridging this science–policy gap is essential for ensuring that advances in renewable technologies are accompanied by viable institutional, regulatory, and financial structures. Based on the patterns identified in this study, several policy-relevant implications emerge.

The predominance of technological and engineering topics (e.g., t_3, t_5, t_10, t_12, t_13, t_15) compared to themes related to economics, governance, and project management (e.g., t_2, t_18) highlights a critical gap in the regional research ecosystem. Policymakers and funding agencies should establish targeted incentives—such as dedicated calls for interdisciplinary projects—that encourage collaboration between engineers, economists, social scientists, and policy experts. This would ensure that technological advances are embedded in feasible business models and supported by robust governance frameworks, increasing their scalability and policy relevance.

Our collaboration network analysis shows that research activity is disproportionately concentrated in Brazil, Mexico, and Chile, leaving many Central American and Caribbean nations underrepresented. To address this imbalance, regional organizations should establish collaborative research consortia and shared funding instruments, modeled on initiatives like the European Union’s Horizon programs. Such frameworks would strengthen cross-border projects, promote knowledge transfer, and build research capacity in underrepresented countries, ultimately fostering a more cohesive and equitable regional research landscape.

The growing prominence of artificial intelligence in renewable energy research (t_7) underscores the potential of data-driven decision-making. However, the integration of these tools into national and regional policy frameworks remains limited. Governments should invest in building open-access data repositories that consolidate information on energy generation, consumption, and meteorological patterns. In parallel, they should establish regulatory frameworks that incentivize the use of AI-driven forecasting and optimization models in national energy planning, investment prioritization, and grid management. This would not only improve efficiency and resilience but also enable evidence-based policymaking aligned with global climate targets. Therefore, by addressing these areas, LAC policymakers and funding agencies could ensure that scientific advances in renewable energy are more effectively translated into inclusive, region-specific policies that accelerate the transition to sustainable and resilient energy systems.

Additionally, the unequal distribution of renewable energy research across Latin America and the Caribbean cannot be explained solely by differences in scientific output; it reflects deeper structural drivers rooted in historical, institutional, and systemic factors. One of the most critical dimensions is the disparity in national R&D funding and infrastructure. Larger economies such as Brazil, Mexico, and Chile have consistently invested more in research and higher education, allowing them to establish advanced laboratories, support graduate programs, and attract international partnerships. In contrast, many Central American and Caribbean nations operate with limited research budgets, restricting their ability to sustain long-term scientific agendas or participate in global knowledge networks. Another key factor is the orientation of international collaboration networks, which often favor partnerships with well-funded institutions in the Global North (e.g., the USA and Europe) rather than fostering intra-regional cooperation. While these collaborations enhance global visibility, they can inadvertently perpetuate dependency on external funding and limit the development of regionally tailored solutions to local energy challenges.

Language and institutional barriers further exacerbate these inequalities. Research in LAC is conducted primarily in Spanish and Portuguese, while high-impact journals predominantly publish in English. This linguistic divide restricts the international dissemination of research from smaller countries and creates asymmetries in visibility and citation impact. In addition, variations in institutional priorities, regulatory frameworks, and educational systems across LAC hinder the development of a cohesive research ecosystem. These structural challenges are compounded by the historical legacies of colonialism, which have shaped uneven patterns of scientific investment, education access, and international recognition across the region. Together, these factors contribute to a fragmented research landscape where a few countries dominate renewable energy scholarship, while others remain underrepresented. Recognizing these underlying drivers is essential for designing policies and funding mechanisms that not only increase research capacity but also reduce systemic inequalities in the scientific ecosystem of LAC.

In this way, research in LAC revealed that disparities in scientific output are deeply intertwined with the region’s broader challenges in renewable energy development, which are rooted in long-standing structural inequalities [46]. Larger economies such as Brazil and Mexico benefit from stronger institutional support and more robust funding mechanisms, which translates to greater research capacity [47]. In contrast, many smaller nations grapple with significant funding constraints and limited infrastructure, hindering their participation in the scientific landscape [48]. These imbalances are exacerbated by underlying socioeconomic factors, including disparities in education systems and resource access, which can marginalize low-income populations and impede an inclusive energy transition. These contemporary inequalities are not new; they reflect a deep historical context shaped by the region’s colonial legacy and persistent social stratification based on class and race, which have been reinforced by underdeveloped state structures [47]. Furthermore, non-economic factors, such as national consumption patterns and the influential role of global finance capital, also contribute to the divergent paths of renewable energy investment and development across the region [46,48].

4.1. Emerging Research Frontiers in a Global Context

The results show a marked increase in renewable energy publications, with a notable acceleration after 2015. This inflection point coincides with the adoption of the Paris Agreement and the Sustainable Development Goals, corroborating similar findings in global bibliometric analyses that associate international policy milestones with surges in scientific output [49,50]. At the country level, the concentration of scientific production in Brazil, Mexico, and Chile parallels patterns observed in the European Union, where Germany, Spain, and the UK dominate output [51].

The thematic trajectory identified through the HJ-Biplot confirms the maturation of the research agenda, evolving from early work on solar radiation modeling and photovoltaic materials toward broader systems-level challenges such as energy storage, microgrids, and decarbonization policies. Similar shifts have been documented in Latin America’s solar and wind research, which moved from foundational assessments to more integrated approaches including economics, maintenance, and grid optimization [52,53]. This transition underscores the increasing recognition that technological advances must be accompanied by institutional, financial, and regulatory frameworks if they are to contribute to sustainable development [54].

Topic modeling identified artificial intelligence (AI) applications (t_7) and microgrid optimization (t_17) as prominent emerging areas. These findings mirror global trends emphasizing digitalization and decentralization in the energy sector. International reviews consistently highlight the transformative potential of AI for demand forecasting, predictive maintenance, and real-time optimization of renewable energy systems [55]. The rapid integration of AI into renewable energy research in LAC therefore signals convergence with international frontiers and suggests that regional scientific capacity is adapting to the digital transition.

The strong presence of microgrid research aligns with the increasing global emphasis on decentralized energy systems as pathways to resilience and inclusivity [56]. For LAC, this trend has particular resonance, given the region’s geographic diversity and the challenges of extending centralized grids to rural and island communities. However, the absence of offshore renewable energy as a distinct topic in LAC stands in contrast to the global literature, where offshore wind, wave, and tidal energy represent rapidly expanding areas of innovation [57]. This divergence suggests that LAC’s research agenda is oriented toward leveraging abundant onshore resources such as hydropower and biomass while postponing more capital-intensive offshore initiatives.

4.1.1. Climate Change Policies and Decarbonization Strategies

One of the most notable trends identified in this study is the increasing research focus on climate change policies, energy transition planning, and decarbonization strategies (topic t_8). This shift underscores the growing recognition of robust regulatory frameworks and policy-driven energy transitions, which are essential for achieving net-zero emissions and enhancing environmental governance. The rise in research in these areas is consistent with global studies emphasizing the importance of strong policy instruments to meet commitments under the Paris Agreement and the 2030 Agenda for Sustainable Development [58,59].

As LAC countries refine their climate-resilient energy strategies, the energy transition must simultaneously address socio-economic disparities, particularly in terms of access to clean and affordable energy. A just energy transition must prioritize energy security and independence while also ensuring that vulnerable populations and underserved regions benefit from sustainable energy development. Future research should explore the role of policy incentives, financing mechanisms, and governance structures in accelerating the transition while mitigating socioeconomic inequalities.

4.1.2. Artificial Intelligence (AI) in Renewable Energy Management

Another significant trend is the expansion of AI applications in energy management (topic t_7), reflecting a shift toward intelligent, data-driven energy systems. The integration of AI technologies has become essential for optimizing energy production, distribution, and consumption, enabling real-time decision-making to improve grid stability and efficiency. AI-driven models can enhance demand forecasting, optimize load balancing, and improve grid automation, ultimately reducing operational costs and mitigating the intermittency of renewable sources such as solar and wind energy.

This trend is further reinforced by the growth of AI-focused startups, research hubs, and technology incubators in LAC, which are driving the development of AI-powered tools for energy optimization. The use of machine learning for predictive maintenance and real-time monitoring also enhances the long-term sustainability of renewable energy systems. As the region moves toward a smarter energy infrastructure, AI-based solutions will play an integral role in ensuring efficiency, reliability, and sustainability in the renewable energy sector [60].

4.1.3. Microgrids and Energy Decentralization

A second key research area in LAC is the increasing emphasis on microgrids and energy optimization (topic t_17). Decentralized energy systems, particularly microgrids, are gaining traction as viable solutions to energy access disparities and grid vulnerabilities, particularly in rural and remote communities. Research in this field highlights the critical role of smart inverters, battery storage, and microgrid architectures in improving the stability and resilience of renewable energy systems. These technologies not only enhance voltage regulation and minimize energy losses but also facilitate the seamless integration of solar and wind power into local grids.

The growing adoption of distributed generation systems across LAC reflects a strategic shift towards localized energy production, reducing reliance on centralized grids that are often fragile in the face of climate-related disruptions. The ability of microgrids to operate independently makes them particularly valuable for disaster-prone regions, where climate resilience is a key consideration. Expanding research on microgrid deployment, energy storage optimization, and policy incentives will be crucial for scaling up decentralized energy infrastructure across the region [61,62].

4.2. Challenges in Regional Research Collaboration

Despite the increasing volume of scientific output, this study highlights a significant polarization in regional research collaboration. Network analysis revealed that a small group of highly connected countries (Brazil, Mexico, Chile) dominate research activity, while many smaller nations remain weakly integrated into regional collaboration networks. Several structural and systemic barriers contribute to this imbalance:

• Unequal research funding and infrastructure: Investment in research and development (R&D) varies greatly across LAC, with some countries allocating substantial resources to scientific research, while others face severe funding constraints. Limited access to cutting-edge laboratories, computing facilities, and field research opportunities restricts the ability of underfunded institutions to engage in high-impact scientific collaborations.
• Limited regional collaboration: While international collaborations (particularly with North American and European institutions) are prevalent, intra-regional scientific partnerships remain weak. Many LAC researchers prefer to collaborate with better-funded global institutions due to higher funding opportunities and greater international visibility, reducing knowledge transfer within the region.
• Language and institutional barriers: Scientific research in LAC is conducted in multiple languages (Spanish, Portuguese, and English), which can create challenges in research dissemination and accessibility. Additionally, differences in institutional policies, research priorities, and regulatory frameworks further hinder the development of cohesive regional collaborations.
• Data accessibility and knowledge gaps: Open-access data is crucial for renewable energy research, yet many LAC countries lack centralized repositories for energy-related datasets. This limits comparative research and data-driven policy development, particularly in emerging fields like smart grids, where real-time data sharing is essential for optimizing energy management.

4.3. Opportunities for Strengthening Regional Collaboration

Despite these challenges, strategic interventions can significantly enhance research collaboration and scientific impact in LAC. Establishing regional research consortia focused on renewable energy can serve as a structured platform for knowledge exchange, joint funding applications, and shared research infrastructure. Inspired by the European Union’s Horizon programs, such initiatives could facilitate cross-border research efforts.

Leveraging AI-driven collaboration tools, digital platforms, and open-access data repositories can also enhance connectivity among researchers, reducing barriers to participation for institutions with limited resources. Expanding funding mechanisms—such as government-backed grants, private sector partnerships, and international research funds—can provide essential financial support for joint projects. Strengthening multidisciplinary and policy-oriented research can ensure that scientific advancements translate into actionable energy policies, while capacity-building initiatives (e.g., researcher exchanges, joint training workshops, and PhD collaborations) can bridge the expertise gap among LAC institutions. The integration of local knowledge with scientific research and policy development is not only beneficial but necessary to foster inclusive energy systems that cater to the diverse needs of LAC’s populations [4].

Additionally, integrating local and indigenous knowledge into renewable energy research can enhance community-driven energy transitions. Many renewable energy projects in LAC operate in regions with indigenous and rural populations, yet these communities remain underrepresented in energy planning. A more inclusive research approach, engaging local populations as co-creators of knowledge, can increase social acceptance and long-term sustainability of renewable energy initiatives.

The collaboration network analysis highlights structural disparities, with Brazil, Mexico, and Chile serving as regional hubs while many smaller nations remain marginal. This asymmetry resonates with bibliometric studies of Latin American science, which repeatedly identify concentration of resources in a few countries and institutions [50,54]. Similar imbalances have been reported in the European Union, though collaborative frameworks such as Horizon Europe have helped mitigate them [51].

The finding that intra-regional collaborations are weaker than international partnerships reflects the “center-periphery” dynamic frequently described in the literature on global science. Studies show that LAC researchers often collaborate more with partners in the United States and Europe than with regional peers, accessing greater funding and visibility but limiting the development of a cohesive, locally driven research ecosystem [49,50]. This dynamic reinforces dependence on external agendas, which may not always align with LAC’s unique socio-environmental challenges.

Taken together, the findings of this study align with broader global bibliometric evidence showing that renewable energy research is increasingly shaped by the dual imperatives of technological innovation and sustainable development [51]. The steady growth of scientific production in LAC reflects the region’s integration into this global movement, with thematic evolution toward policy, decarbonization, and digitalization echoing trends documented worldwide. Importantly, these trajectories directly support the achievement of Sustainable Development Goals 7 and 13, confirming the central role of renewable energy innovation in advancing the global sustainability agenda. At the same time, the analysis underscores persistent structural imbalances within the regional research ecosystem. The dominance of Brazil, Mexico, and Chile mirrors patterns of concentration observed in other regions such as the European Union [51], while the relative absence of research output from smaller Central American and Caribbean nations reflects deeper systemic inequalities rooted in funding disparities, institutional capacity, and limited regional collaboration [56,57]. Addressing these disparities is critical to ensure that the benefits of renewable energy transitions are equitably distributed and that regional research agendas adequately reflect the diversity of local contexts. By highlighting both the convergences and divergences between LAC and global research trends, this study provides a roadmap for strengthening regional scientific collaboration, aligning research priorities with sustainability targets, and advancing inclusive pathways toward a low-carbon future.

4.4. Policy Implications and Research Gaps

While research output in renewable energy is expanding, its translation into policy and practical implementation remains a challenge. This study identifies a disconnect between scientific research and policy formulation in LAC, suggesting that greater efforts are needed to ensure that research findings inform national and regional energy policies.

Governments and energy agencies in LAC can leverage the findings from this study to design policies that facilitate the integration of renewable energy. Some key areas where research can directly inform policy include:

• Developing targeted financial incentives to support the adoption of renewable technologies.
• Strengthening regulatory frameworks for microgrid integration and distributed energy generation.
• Enhancing regional cooperation for data-sharing and energy infrastructure development.

This study also reveals gaps in research related to the economic feasibility of renewable energy projects, social acceptance of new technologies, and the long-term impacts of energy transition policies. Addressing these gaps through interdisciplinary research will be crucial for achieving a just and equitable energy transition.

Although AI is being increasingly used for technical optimization, its potential for informing policy decisions remains underexplored. Future research should examine how AI-based forecasting models can be utilized for national energy planning, policy evaluation, and investment prioritization. The development of AI-driven policy recommendation tools could help governments make more informed decisions on energy transition pathways.

AI has the potential to revolutionize energy governance by enhancing predictive analytics for renewable energy supply and demand fluctuations. Governments and policymakers could leverage machine learning algorithms to optimize subsidy allocations, assess the long-term impact of policy interventions, and improve energy security. Additionally, AI-driven simulations can support scenario planning, allowing decision-makers to evaluate various energy transition pathways and their socio-economic implications.

Furthermore, AI can facilitate international cooperation by enabling real-time data exchange and harmonization of policy frameworks across borders. Through AI-powered insights, countries in LAC can enhance collaboration on energy infrastructure development, integrate regional renewable energy grids, and develop shared investment strategies. Future research should focus on refining AI tools for multi-criteria decision analysis, which can aid in balancing economic growth, environmental sustainability, and social equity in renewable energy policies. Although AI is being increasingly used for technical optimization, its potential for informing policy decisions remains underexplored. Future research should examine how AI-based forecasting models can be utilized for national energy planning, policy evaluation, and investment prioritization. The development of AI-driven policy recommendation tools could help governments make more informed decisions on energy transition pathways

4.5. Contribution to the Sustainable Development Goals (SDGs)

The findings of this study have significant implications for the advancement of the United Nations Sustainable Development Goals (SDGs), particularly SDG 7 (Affordable and Clean Energy) and SDG 13 (Climate Action). The sustained growth in renewable energy research in LAC reflects an increasing commitment to promoting clean energy solutions, improving energy access, and mitigating the environmental impact of energy production.

SDG 7 emphasizes the need to ensure universal access to affordable, reliable, and modern energy services. The research trends identified in this study demonstrate a regional shift toward developing and optimizing renewable energy technologies such as solar, wind, biomass, and microgrid systems. The increasing focus on decentralized energy systems, particularly microgrids, directly supports SDG 7 by facilitating energy access in remote and underserved communities. Additionally, the integration of AI and advanced predictive analytics in energy management contributes to enhancing the efficiency and reliability of renewable energy systems, making clean energy more accessible and sustainable across LAC.

SDG 13 calls for urgent action to combat climate change and its impacts. The expansion of research on climate change policies, decarbonization strategies, and energy transition planning indicates that LAC is aligning its research efforts with global climate objectives. The increasing body of work on carbon emission reduction, energy storage solutions, and policy-driven renewable energy integration contributes to the region’s ability to implement science-based strategies for reducing greenhouse gas emissions. Furthermore, the growing application of AI in renewable energy optimization enhances climate resilience by improving forecasting accuracy and facilitating real-time adjustments to energy supply and demand, reducing dependency on fossil fuels.

By bridging the gap between scientific research, policy development, and technological innovation, this study underscores the critical role of renewable energy research in supporting sustainable development efforts in LAC. The integration of cutting-edge technologies, such as AI-driven energy solutions, along with strong policy frameworks, will be essential for ensuring a just and effective transition toward a low-carbon energy future.

4.6. Study Limitations and Future Research Directions

This study offers a comprehensive overview of renewable energy research trends in LAC, yet certain limitations should be acknowledged to contextualize the findings. The analysis is limited to peer-reviewed publications indexed in Scopus and Web of Science, which may exclude relevant grey literature such as government reports, industry white papers, and NGO studies. As a result, the representation of policy-driven initiatives and practical implementation may be incomplete. Moreover, the reliance on bibliometric and topic modeling methods makes this study descriptive and correlational rather than causal. While the analysis identifies significant patterns and connects them with external events, it does not formally test the causal mechanisms that drive these observed shifts.

It is important to recognize that the methodological scope of this study is descriptive and correlational rather than causal. While our findings highlight strong temporal trends and their alignment with major policy developments, the bibliometric and topic modeling approach does not allow for the formal testing of causal mechanisms (e.g., proving that a specific policy directly caused a shift in research priorities). Future studies could build on our results by employing complementary methodologies, such as qualitative case studies, econometric analyses, or mixed-methods approaches, to rigorously investigate the causal drivers behind the thematic and collaborative dynamics identified here.

Future research can build upon these limitations in several directions. First, qualitative case studies and econometric models could provide stronger causal insights, for example by examining the effect of specific national funding initiatives, education policies, or regulatory reforms on research productivity and thematic focus. Second, systematic reviews of grey literature would help bridge the gap between scientific research and real-world implementation, revealing the extent to which academic priorities align, or diverge, from policy frameworks and industrial practices. Third, collaboration dynamics could be explored in greater depth by moving beyond co-authorship networks to investigate flows of funding, institutional partnerships, and the diffusion of knowledge across borders. Such approaches would enrich understanding of how research capacity and structural barriers shape regional scientific development.

As the region advances toward a sustainable energy transition, these future lines of inquiry will be essential for connecting academic discovery with practical action. Strengthening interdisciplinary research, expanding intra-regional collaboration, and closing the gap between science, policy, and industry will be critical to ensuring that renewable energy development in LAC is inclusive, resilient, and aligned with global sustainability goals.

This study offers a comprehensive overview of renewable energy research trends in LAC, yet several limitations should be acknowledged to contextualize the findings. First, the analysis was restricted to peer-reviewed publications indexed in Scopus and Web of Science, which may exclude relevant grey literature such as government reports, industry white papers, NGO studies, and regional journals not covered by international databases. As a result, the representation of policy-driven initiatives, community-based projects, and practical implementation experiences may be incomplete. Second, the reliance on bibliometric and machine learning techniques, including LDA, MDS, and HJ-Biplot, while powerful for detecting large-scale trends, is inherently descriptive and correlational. Although this study identifies significant patterns and situates them within global policy milestones such as the Paris Agreement and the SDGs, it does not formally test the causal mechanisms driving these shifts. Finally, the analysis was limited to articles and reviews, leaving aside other forms of scientific production such as conference proceedings, technical reports, and governmental policy documents that may capture more applied or emerging insights.

Future research can build upon these limitations in several directions. Qualitative case studies and econometric models could provide stronger causal insights, for example by assessing the influence of national funding programs, education policies, or regulatory reforms on research productivity and thematic focus. Systematic reviews of grey literature would help bridge the gap between academic research and practical implementation, clarifying the extent to which scholarly priorities align, or diverge, from policy frameworks and industrial practices. Expanding collaboration analyses to include funding flows, institutional partnerships, and knowledge diffusion mechanisms could reveal structural drivers of inequality in research capacity across the region. Additionally, incorporating local and indigenous knowledge, community participation, and governance structures into future studies would enrich the understanding of how renewable energy adoption unfolds in diverse socio-political contexts. From a methodological perspective, hybrid approaches that combine topic modeling with simulation techniques, such as agent-based or network-based models, may yield deeper insights into the dynamics of collaboration, policy diffusion, and technology transfer. By addressing these limitations, future research can strengthen the link between scientific evidence, policy development, and practical action. Interdisciplinary approaches, expanded intra-regional collaboration, and closer integration of science, policy, and industry will be critical to ensuring that renewable energy development in LAC is not only technologically advanced but also socially inclusive, resilient, and aligned with global sustainability goals.

5. Final Remarks and Conclusions

This study provides a comprehensive analysis of the evolution of scientific production on renewable energy in LAC over the past three decades, identifying key research trends, thematic priorities, and methodological advancements. By leveraging machine learning-based topic modeling and bibliometric techniques, this research offers novel insights into the structure, development, and trajectory of renewable energy studies in the region. The findings contribute to a data-driven understanding of which areas have received the most attention, how research priorities have shifted over time, and where critical gaps remain.

A central conclusion of this study is the steady growth in renewable energy research output, reflecting increasing regional and global efforts to transition toward low-carbon and sustainable energy systems. This expansion aligns with international climate commitments, particularly the Paris Agreement and the Sustainable Development Goals (SDGs 7 and 13). However, the research landscape remains unevenly distributed, with Brazil, Mexico, Colombia, and Chile leading scientific output, while smaller Caribbean and Central American nations remain significantly underrepresented. This disparity underscores the urgent need for greater regional collaboration, research funding, and policy support to bridge the knowledge gap and promote research initiatives across all LAC nations.

One of the key methodological contributions of this study is the application of machine learning techniques, specifically Latent Dirichlet Allocation (LDA) topic modeling, to identify emerging research themes and evolving priorities. This approach enhances the detection of latent patterns in scientific literature, providing a structured, scalable, and automated alternative to conventional literature reviews. The findings demonstrate that AI and data science have the potential to revolutionize the way scientific knowledge is analyzed, predict research trends, and support evidence-based policymaking in the renewable energy sector.

The results highlight the increasing role of policy-driven research, with a growing number of studies focused on climate change mitigation, energy transition frameworks, and decarbonization strategies. This trend aligns with global efforts to accelerate the deployment of renewable energy technologies and reduce fossil fuel dependency. Additionally, this study identifies a notable rise in microgrid and energy optimization research, reflecting a shift toward decentralized energy systems that enhance grid resilience, efficiency, and accessibility, particularly for rural and off-grid communities. This transition is crucial for reducing energy inequality in LAC, where many isolated regions still lack reliable electricity access.

Another key finding is the growing integration of AI in renewable energy research. AI-driven models are increasingly used for demand forecasting, energy storage optimization, and smart grid management, reflecting a broader trend toward digital transformation in the energy sector. The emergence of AI-based startups, research initiatives, and technological partnerships suggests that intelligent energy management systems will play an essential role in improving the efficiency, reliability, and scalability of renewable energy solutions.

Despite these advancements, this study also identifies persistent gaps and challenges that need to be addressed to further strengthen renewable energy research in LAC. Some critical areas remain underexplored, including economic feasibility assessments, regulatory frameworks, and socio-environmental impacts of renewable energy adoption. These findings suggest that interdisciplinary research approaches—integrating engineering, economics, social sciences, and environmental studies—are needed to develop more holistic and inclusive energy solutions. Furthermore, limited international and intra-regional collaboration presents a barrier to knowledge transfer and innovation, highlighting the need for stronger research networks, joint funding mechanisms, and cross-border initiatives that facilitate scientific cooperation and technological advancements.

To maximize the impact of renewable energy research in LAC, it is essential to promote evidence-based policymaking, increase investment in research and development (R&D), and foster stronger partnerships between academia, industry, and government stakeholders. Ensuring that research findings translate into actionable policies and practical solutions will be crucial for accelerating the energy transition. Additionally, expanding the adoption of machine learning, big data analytics, and AI-driven modeling can enhance scientific discovery, optimize decision-making, and improve the efficiency of renewable energy systems.

In conclusion, this study provides a solid foundation for understanding the trajectory of renewable energy research in LAC, offering valuable insights for policymakers, researchers, and industry leaders. By applying machine learning techniques to bibliometric data, this research contributes to a more systematic, scalable, and data-driven approach to analyzing scientific production and research evolution. The findings emphasize the crucial role of digitalization, decentralization, and policy integration in shaping the future of renewable energy in the region. Moving forward, fostering collaborative research efforts, supporting emerging technologies, and addressing regional disparities will be essential to achieving a just, inclusive, and sustainable energy transition in Latin America and the Caribbean.

Finally, this study underscores the critical importance of scientific collaboration in driving the renewable energy transition in LAC. While significant progress has been made, regional disparities in research capacity, funding, and connectivity must be addressed to ensure a more inclusive and effective approach to sustainable energy development. Strengthening intra-regional collaboration, fostering interdisciplinary research, and leveraging emerging technologies will be fundamental in positioning LAC as a leader in global renewable energy innovation.