Digital Transformation—One Step Further to a Sustainable Economy: The Bibliometric Analysis

Abstract

Digitalization has significantly reshaped human and social life worldwide, serving as a powerful enabler of a sustainable economy, while being directly aligned with Sustainable Development Goal 9, among others. The literature on digitalization and sustainability boosted since 2017, confirming its importance. Unlike most previous studies, this paper extracted articles from both the Scopus and Web of Science platforms, and the bibliometric analysis was conducted using the new Python library, pyBibX, for the cleaned concatenated dataset, as well as Bibliometrix in R for the parallel analysis on the two platforms. We conducted both a performance analysis to measure scientific impact and citations in the quest to better understand the research field and also a science mapping to visually represent the scientific research and its development. Our findings suggest that Sustainability is the main journal with published articles on digitalization and sustainability, whereas China has the largest number of papers in the field and collaborations between countries. Finally, by applying Natural Language Processing, we identified as best topics: digital, sustainable, development, sustainability, digitalization, study, research, transformation, innovation, and model. Moreover, we dug deeper into policy implications to show how these findings could serve policymakers and stakeholders in academia and industry.

1. Introduction

Since the Fourth Industrial Revolution, when notable technological advancements were achieved, corroborated with new digital job skills requirements, information and communication technology has emerged as a key driver of transformation in business and organizational practices. Recent advances in technology have led to innovations in all economic sectors around the world.

In terms of emerging technologies, two primary dimensions were identified by Brynjolfsson & McAfee [1]. The first one refers to increased machine power, encompassing technologies such as artificial intelligence (AI), big data, augmented reality, advanced robotics, autonomous vehicles, and 3D printing. The second dimension involves enhanced connectivity, which includes technologies such as the mobile Internet, social media, audio and video conferencing, the Internet of Things (IoT), cloud computing, and blockchain.

With increasing enthusiasm in the digital field, the terms “digitization” and “digitalization” are frequently used interchangeably in academic research. Digitization encompasses the multitude of transformations that occurred with the Third Industrial Revolution and represents the conversion of real-life information into its digital (binary) form, allowing for the acquisition of information across multiple platforms [2]. Digitalization, on the other hand, is a much more extensive concept, as it marks the beginning of the Fourth Industrial Revolution [3]. This notion introduces the novelty of reconstructing socioeconomic domains around preexisting digital changes. In other words, digitalization is the act of converting the digital form created by digitization into meaningful information that can be used to optimize the evolution of society. The digital transformation concept expands on the two previously outlined notions by specifically representing the organizational area through digital changes implemented by enterprises in a variety of ways [4]. As the most recent stage of technological evolution is digitalization, it is noted that it has significantly reshaped human and social life [5]. This technological shift has fundamentally transformed traditional paradigms, driving them towards digital forms.

The importance of digitalization has been emphasized by the European Commission through its policy framework. For instance, the “Digital Agenda for Europe” [6] supports Europe’s digital transformation from multiple angles. In terms of job creation and skills development, it encourages the growth of start-ups and small- and medium-sized enterprises (SMEs) in the tech sector, generating new job opportunities and advancing digital training. To enhance competitiveness, the initiative promotes the development of e-government services, improving the efficiency and accessibility of public administration, and drives the digitalization of industries to become more technology-driven and data-focused. It also emphasizes the need for improved broadband infrastructure and connectivity across Europe.

Moreover, the “A Europe Fit for the Digital Age” initiative [7] advocates for digital transformation and innovation, emphasizing the importance of cybersecurity and a green digital transition. It encourages investment in digital education and the adoption of technologies such as AI, blockchain, and high-performance computing to boost productivity and competitiveness across industries. These efforts have opened up new opportunities for companies to succeed in the digital economy while fostering the development of digital skills, creating jobs, and driving economic growth in the digital sector [8].

The crucial role of digitalization in achieving the Sustainable Development Goals (SDGs) is also emphasized through the 2030 Agenda for Sustainable Development established by the United Nations in 2015. Comprising 17 interconnected goals, the SDGs address a broad range of global challenges, including inequality, climate change, digitalization, circular economy, smart cities, peace, and justice. Specific targets and indicators correspond to each goal to help guide countries to reach economic growth, social inclusion, and environmental sustainability. The essence of the SDGs lies in their holistic approach, highlighting the importance of collaboration among governments, businesses, academia, and individuals to build a sustainable future that leaves no one behind.

Digitalization aligns most directly with SDG 9, which supports building resilient infrastructure, promotes inclusive and sustainable industrialization, and fosters innovation. However, the cross-cutting nature of digitalization allows supporting broader sustainability objectives across multiple SDGs. It serves as a powerful enabler of a sustainable economy, as it optimizes resource allocation and reduces carbon emissions through technologies such as IoT, AI, and data analytics. Digital platforms also enable sustainable business models, such as product-as-a-service and the sharing economy, supporting green innovation. Additionally, digitalization enhances environmental monitoring, data-driven decision making, and climate resilience, helping societies and businesses adapt to sustainability challenges more effectively [9].

Having these in consideration, this paper aims to determine the most representative papers, authors, and topics for the domain of digitalization and sustainability. Unlike previous studies that mostly explored documents extracted from just a single platform such as Web of Science (WoS) or Scopus, the novelty of this research comes from the fact that it conducts a more comprehensive and objective bibliometric analysis on the relationship between digitalization and sustainability by gathering information from two of the most scientifically well-known databases (WoS and Scopus) and performs the analysis on a combined and cleaned set of data using the new Python library, pyBibX. For comparison reasons with previous literature, we also conduct a parallel analysis of the documents from each of the two databases. Our bibliometric analysis consists of two steps: (1) firstly, a performance analysis is conducted to measure scientific impact and citations through several indexes to better understand the research field in terms of structure and trends in academic activity; (2) secondly, a science mapping is performed to visually represent the scientific research and its development in the academic and social fields. Moreover, we dig deeper into the policy implications of our research to argue how our bibliometric analysis findings could serve policymakers and stakeholders in academia and industry. The main outcomes will be presented as answers to the following research questions that we raised:

RQ1: How has research on digitalization and sustainability been distributed between 2017 and 2024 in terms of publications and citations?

Through this research question, we intend to evaluate the current trend of publication on the nexus of digitalization and sustainability. Since the outburst of the COVID-19 pandemic, sustainable development was at risk worldwide, and digitalization emerged as a means to address economic, social, and ecological crises. Thus, academic activity extended especially on topics concerning digital transformation and ways to foster innovation and to achieve SDGs through digitalization.

RQ2: What are the most influential articles, authors, journals, and countries in this field?

For this research question, we intend to conduct a performance analysis to better understand the research field in terms of structure and trends in academic activity and to compare our results to previous studies in the field.

RQ3: What are the collaboration trends?

By this research question, we plan to identify collaboration clusters between authors, as well as the most productive country, in terms of published papers on the topic of digitalization and sustainability.

RQ4: How has the research output evolved over time?

Based on this research question, we intend to identify both the hot topics and the emerging themes in the field of digitalization and sustainability. These findings could guide both governmental and corporate decision-makers to prioritize areas for development and further exploration. Policymakers could thus plan their sustainability strategies more effectively and create a virtuous circle for stakeholders. These digital technologies have the potential to drive the development of environmentally friendly business models while enhancing profitability through the digital transformation of production processes.

RQ5: What are the main topics in the literature on digitalization and economic sustainability?

Finally, we intend to analyze the content and results of the selected studies in order to categorize the main streams of research from existing studies on digitalization and sustainability and to investigate the role of digitalization in sustainable development. Two notions that are connected but concentrate on two different characteristics are encountered in the field of sustainability. In study findings, the phrase “sustainable development” emphasizes sustainable features at the global economy, environmental, and social levels, whereas the phrase “sustainable economy” exclusively concentrates on economic factors. Therefore, the essential difference between the two lies in the scale considered: the extended global level provided by “sustainable development” and the compressed level of economic systems provided by “sustainable economy” [10,11].

Our findings will help detect research gaps and formulate recommendations for future investigation. The structure of the paper is the following. Section 2 serves as a brief literature review on the topic of digitalization and sustainability mostly investigated through bibliometric analysis. Section 3 describes the methodology, including the datasets, the bibliometric indicators, and the tools and software used in the analysis, while Section 4 presents the main results of our bibliometric analysis conducted on the nexus between digitalization and sustainability. Finally, Section 5 discusses the contributions to theoretical research by connecting findings to RQ, along with their policy implications, while the last Section 6 concludes.

2. Literature Review

2.1. Digitalization and Economic Sustainability

Digitalization plays a crucial role in driving sustainable economic development and achieving the SDGs. Firstly, through the integration of digital technologies into various sectors, it enhances sustainable industrialization, productivity, and innovation, all of which are essential for achieving SDG 9. To support this, Nayal et al. [12] argue that technologies such as AI, IoT, and blockchain lead to a circular economy, which is crucial for sustainable industrial practices. In the context of Industry 4.0, these advanced technologies help SMEs adapt to changing market conditions and optimize production processes, while minimizing environmental impacts [13]. This perspective is also supported by Velden [14], who states that digital technologies can regenerate the planet by fostering sustainable practices across various sectors. Moreover, Łobejko & Bartczak [15] state that digital technology platforms facilitate sustainable consumption and production by promoting sharing economies and optimizing production processes. Furthermore, digitalization enhances the quality of services and productivity in SMEs, which are often the backbone of economies [16]. Additionally, Ufua et al. [17] argue that fostering collaboration among stakeholders can lead to innovation and more sustainable outcomes. This aligns with the findings of Ordieres-Meré et al. [18], who argue that intra-organizational knowledge sharing through digitalization leads to sustainable outcomes in manufacturing (SDGs 2 and 12).

Secondly, digital transformation can effectively support the achievement of SDGs by creating new opportunities for economic growth and sustainability [19]. This is also supported by Gupta et al. [20], who argue that digitalization reduces resource consumption and greenhouse gas emissions while also promoting education and welfare. The ability of digital technologies to streamline processes and improve resource management is essential for achieving the SDGs related to economic growth (SDG 8) and responsible consumption (SDG 12).

Moreover, digitalization plays a notable role in enhancing food security and sustainable agricultural practice. Vărzaru [21] discusses how digital transformation can improve food security and related SDGs within the European Union, emphasizing the link between digitalization and sustainability in food systems. In addition to operational improvements, digitalization also helps promoting financial inclusion, which is vital for reducing inequalities (SDG 10). Chatterjee [22] states that financial technology is instrumental in supporting balanced and sustainable development by enhancing access to financial services. This is particularly relevant in regions like East Africa, where Koomson et al. [23] highlight the transformative impact of mobile money on entrepreneurship and economic resilience. By enabling access to digital financial services, digitalization can empower marginalized communities and support the achievement of SDGs.

Finally, the global community’s response to the COVID-19 pandemic has underscored the importance of digital transformation in achieving the SDGs. Mospan [24] highlights that while the pandemic brought risks to sustainable development, it also accelerated the digital transformation in higher education. In conclusion, digitalization is a transformative force that not only enhances economic efficiency but also aligns with the principles of sustainability. By creating new opportunities, promoting financial inclusion, and optimizing resource management, digital technologies play a critical role in advancing the SDGs.

The current level of research on the implications of digitalization in the sustainable development of the economy indicates an increased interest in this field. Research on digitalization largely focuses on its impact on the business environment [25], digital transformation [26], the transformation of human resource practices [27], and on how digitalization promotes long-term economic sustainability and green innovation [28,29]. The literature offers various perspectives on digital transformation, which is broadly defined as the convergence of personal and business IT environments and the transformative impact of emerging digital technologies—such as social media, mobile platforms, analytics, cloud computing, and IoT [25].

Regarding the impact of digital policies on green innovations, Luo et al. [28] provide evidence for China that the digital advance of the economy significantly potentiates green innovations at the urban level. This outcome is achieved primarily through indirect means, indicating a spillover effect. The main policies explored by the authors include increasing economic openness, optimizing the industrial structure, and enhancing market potential.

A similar study was extended at the level of the European Union (EU), looking for the influence of the Digital Economy and Society Index (DESI) on sustainability indicators [30]. The investigation proceeded with an assessment of the “Human Capital”, “Connectivity”, “Integration of Digital Technology”, and “Digital Public Services” dimensions of the DESI index. As a result, it was determined that depending on the dimension of the index, the influence might be positive or negative. Furthermore, the indicators that assess sustainability are only marginally affected by the dimensions’ indicators. Although certain indicators exhibit a significant influence on the trajectory of the green economy, other indicators have essentially no impact at all.

Recently, the effect of the development of the digital economy on sustainability was studied by Li et al. in 2021 [31]. The paper analyzed 190 countries between 2005 and 2016, in the context where previous studies demonstrated the fact that human activities after the industrial revolution greatly increased greenhouse gas emissions. The study focused on determining the relationship between the digital economy and CO₂ emissions, with the findings revealing that the expansion of digitalization increases the amount of emissions over time. However, once a stabilization point is reached, the amount of emissions can be mitigated. The authors’ recommendation encourages governments to implement economic digitalization strategies in order to reduce the pollution time caused by the gradual digitalization of each country.

2.2. Bibliometric Analysis in Research

Recently, there has been a growing interest in studying the development of digitalization worldwide through the use of bibliometric analysis. The term “bibliometric analysis” implies the use of quantitative analysis for bibliometric data, such as citations and units of publication [32]. The research focus was most intensively upon the impact of digitalization in the business sector, where only companies with openness towards hi-tech advancements can now survive the latest challenges on the market [33,34].

Previous studies have used bibliometric analysis to explore key topics in digitalization, with a focus on the following scientific research trends: (1) the development of SMEs digitalization; (2) digitalization in local government; (3) the digital Human Resource (HR) transformation and education; and (4) the link between digitalization and sustainable economy.

Bibliometric analysis studies on digitalization in the business sector provide a comprehensive overview of the current research landscape on enterprise digital transformation. For instance, Chen and Shen [35] used HistCite and CiteSpace and found the journal Technology Forecasting and Social Change and author Parida V to be the most prolific sources on enterprise digital transformation, with most publications coming from the USA. Three distinct stages were highlighted based on the citation mapping: first, the budding stage (before 2014), followed by the system formation stage (2015–2019), and finally the diversified development stage (2020–present). The keyword co-occurrence analysis highlighted four main aspects for the enterprise digital transformation conceptual framework, namely, digital technology adoption, digital transformation performance, digital innovation orientation, and also digital dynamic capabilities.

Sarango-Lalangui et al. (2023) [36] also explored the development and trends in scientific research on the digitization of SMEs using R Bibliometric on 650 publications indexed in the Web of Science platform. They confirmed the significantly increased trend in publications, with a primary focus of SME digitalization research on information technologies, digital transformation, industry 4.0, innovation, sustainable development, and public policy. Furthermore, the COVID-19 pandemic has been observed to have accelerated the digital transformation in many SMEs, especially focused on digital tools and technologies.

Holand et al. [37] proposed a taxonomy to distinguish between various digitalization concepts like digitalization, digitization, and digital transformation and discussed their impact on cost reduction, connectivity, and value creation. Using VOSviewer software on a set of 197 publications extracted from Web of Science, they also found that although the number of scientific papers in the field increased exponentially in recent years, the majority of the publications belonged to lower-ranked journals.

Regarding the research on digital transformation in start-ups, Sreenivasan and Suresh (2023) [38] provided an overview using the Biblioshiny package on a set of articles from the Dimension database published between 2016 and 2022. They found that the Sustainability journal has the highest network of other citations in the field. Even though the research on digital transformation in start-ups is still emerging, there is also a notable steep increase in interest since 2017.

Prijanto (2024) [39] studied the literature on digital innovation in accounting and financial management from 1985 to 2023 using VOSviewer. The database consisted of 980 papers extracted from PubMed, IEEE Xplore, and Google Scholar to assure interdisciplinary coverage. Among the major research cluster themes identified, the analysis revealed topics such as big data, auditing, digital finance, and digital transformation. In terms of research trends, the focus changed from strategic issues towards recent developments covering big data, artificial intelligence, and the pandemic. The role of big data, artificial intelligence, and digital transformation in the business world was also emphasized through the top cited articles in the field.

Lam, Lam and Lee (2023) [40] conducted a bibliometric analysis of the digital twin in the supply chain. The concept of a digital twin refers to a digital representation of a physical entity, crucial for driving Industry 4.0. The authors found several key research groups that focused on the design and integration of the digital twin model, as well as the application of the digital twin in quality control. In terms of research trends, the digital twin was first tested in the production line, gaining more importance along with Industry 4.0, as the Internet of Things, big data, machine learning, blockchain, and cloud-based systems extended the digital twin models. Finally, in the current research stage, publications focus on the integration of deep learning, artificial intelligence, and data models for digitalization.

In terms of digitalization in local governments, the holistic review of research publications indicates China, United Kingdom, and the USA as the most notable regions of publication, while the most cited authors belong to China and the Netherlands. Moreover, based on a set of 57 articles extracted from the Scopus database, the most widely published and cited journal source was Sustainability [41].

Regarding research trends in digital HR transformation in education, studies confirm a steep increase in publications in the last five years, mostly from Germany, Spain, the United States, and Russia [24]. These findings can be explained by the adoption of online education that was accelerated by the COVID-19 pandemic and due to the rapid advancement of digital technologies worldwide.

The link between digitalization and sustainable economy was recently investigated through bibliometric analysis by several studies. Some have investigated the smart cities’ contribution to urban sustainability. For instance, Wu et al. [42] used mapping of the knowledge domain to review 965 studies in the field. Their findings reveal the following four main topic clusters: (1) information technology; (2) energy and environment; (3) urban transportation and mobility; and (4) urban policy and development planning. Chen et al. [43] reviewed the literature on the overall impact of digitalization on environmental sustainability in the context of manufacturing. They found a positive contribution of digitalization to environmental sustainability, as the integration of new digital technologies can boost resource efficiency, while the negative environmental burden is mostly due to increased resource and energy use. Moreover, Feroz et al. [44] outlined four key areas of digital transformations including pollution control, waste management, sustainable production, and urban sustainability. Additionally, Castro et al. [45] conducted a systematic overall review of the literature on SDGs and their relationship to digitization. They highlighted several research gaps, including a limited understanding of SDGs complexities and interconnections, design flaws and imbalances, challenges in implementation and governance, inadequate indicators and assessment methods, incomplete adoption and misaligned progress, unclear accountability, and insufficient coordination. Finally, they showed growing expectations of digitalization’s role in achieving the SDGs, particularly through the use of novel data sources, improved analytical capabilities, and the development of collaborative digital ecosystems.

Other studies have focused, however, on the nexus between digitalization and sustainability. For instance, Davidescu et al. [46] used a total cumulative set of 1269 documents indexed in either WoS or Scopus to analyze comparatively the two databases. The main research topics highlighted by the research referred to digitalization, sustainability, digital transformation, the construction industry, the COVID-19 pandemic, and innovation. Beier G. was found to be among the most relevant authors by both WoS and Scopus platforms, while the Sustainability journal was among the most relevant journals to publish on this cross-cutting topic.

Irajifar et al. [47], on the other hand, only studied the literature from WoS regarding the intersection of sustainability and digitalization. The VOSviewer software was used to visualize and map the literature and to identify research topics through co-occurrence analysis. Four major cluster topics were identified, namely, (1) governance, planning, and policy making; (2) energy, emission, consumption, and production; (3) innovation, economy, green, and environment; and (4) systems, networks, Industry 4.0, and supply chain.

Finally, Aleksy [48] studied the relationship between digital transformation and sustainable development in 9527 documents indexed in Scopus for the period 1990–2023. They found five stages of evolutionary development of the issues of digitalization and sustainable development: (1) stage 1 until 2014, when most publications studied general issues on strategic development of industry; (2) stage 2 from 2014 to 2016, with focus on the terms “information and communication” and “economic and social effects”, among others; (3) stage 3 from 2016 to 2018, with interest shifting towards “data mining”, “information management”, and “sustainable development”, among others; (4) stage 4 from 2018 to 2020, focused on terms like “big data”, “cloud computing”, and “green computing”; and (5) stage 5 from 2020, with focus on terms like “digital transformation”, “IoT”, “AI”, “blockchain”, “machine learning”, “digital technologies”, “digital economy”, “circular economy”, “smart city”, and “Industry 4.0”.

Concluding from previous studies, the systematic review analyses on the nexus between digitalization and sustainability are still in their early stages, and further investigation on research and policy implications is required to fill the gaps in the field. Given the above, our research aims to conduct a comprehensive and objective bibliometric analysis of the relationship between digitalization and sustainability, bringing several elements of novelty to this research area. Compared to the previous limited number of studies that mostly performed their analysis on documents extracted from a single platform (in general, WoS or Scopus), or made only a comparative analysis between two databases, we used the information collected from both most well-known and scientifically recognized databases (WoS and Scopus) and conducted both a parallel analysis of the documents from each of the two databases for comparison reasons with previous literature, but also a more comprehensive analysis on the combined and cleaned set of data using the new Python library, pyBibX. Moreover, we look deeper into the policy implications of our research to argue how our bibliometric analysis findings could serve policymakers and stakeholders in academia and industry.

In the literature, we found papers that dealt with parallel analysis ([32], with data from the Scopus and WoS databases, or [49] on PubMed, Scopus, and Web of Science) or a combined dataset ([50] on Web of Science and Scopus). We did not find relevant papers on a mixed approach like ours: a parallel analysis for determining if the results depend on the platform or are supported by the findings in both databases—for example, sources’ production over time or impact by H-index—and a combined analysis for obtaining a larger and, thus, more relevant set of data, especially needed for ML algorithms that require more data in order to be better trained and perform better (for example, the NLP used for topic clustering).

3. Methodology

The essential steps in the development of the bibliometric analysis, according to Cobo et al. [51], can be seen in Figure 1. The analysis begins with the extraction of the database based on the terms relevant to digitalization and sustainability. The next step consists of analyzing the most important aspects specific to the database: general information, sources, authors, contributing countries, keywords, and the most cited documents. Following the extracted information, the discussions, conclusions, and limitations of the thesis are formulated.

3.1. Data Sources

Web of Science (owned by Clarivate) and Scopus (produced by the Elsevier Company) are two of the most well-known and scientifically recognized indexing databases, along with Google Scholar, Microsoft Academic, and PubMed. In many countries, the scientific development of academia is evaluated based on indexing criteria (H-index, number of citations per paper, number of ISI indexed papers, etc.). Because PubMed has narrower scope and coverage [49], being specific to biomedical literature (which is not our interest); Google Scholar does not have a reliable mechanism for author attribution, being an unpaid service; and Microsoft Academic was closed in 2022, we chose to export data from Scopus and WoS only.

Most reviews and articles extract the data from one of these databases because of issues regarding the exporting formats and in order to avoid duplicates. Previous studies focused on the Scopus database explored the environmental, social, and governance research progress [52], while others used VOSviewer and Bibliometrix to analyze the impact of investment for sustainable development [53]. Another approach identified in the recent studies includes mapping the landscape of misinformation detection [54] by using WoS as source. The parallel analysis of sustainable development and Industry 4.0 was also a main focus in the research field [55], with the alternative of analyzing digital technologies for triple bottom line sustainability [56] from a bibliometric point of view.

pyBibX is a Python library developed and maintained by the Python community and shared by Professor Valdecy Pereira from Universidade Federal Fluminense in Brazil on GitHub. pyBibX reads data from .bib files (Scopus or WoS) and .txt files (PubMed).

Bibliometrix accepts input formats such as .bib (from WoS or Scopus databases), .txt (from WoS, Scopus, PubMed, and Cochrane Library databases), .ciw (from the WoS database), .csv (from Dimensions or The Lens databases), and .xlsx (from the Dimensions database).

In WoS, data can be exported in different formats, such as text, RIS, BibTex, Excel, Tab delimited file, etc., but no more than 500 records at a time if the full record and cited references are exported.

In Scopus, data can be exported as CSV, text, RIS, and BibTex files, including citation information, bibliographical information, abstract and keywords, funding details, conference information, or references. Up to 20,000 documents can be downloaded at a time, regardless of the exporting format.

For this research, we exported information about 1885 documents in Scopus and 1426 documents in WoS (in 3 steps of 500, 500, and 426 rows, put together in a final document). We performed a parallel analysis in Bibliometrix and a combined analysis in Python, using the pyBibX package. This approach allowed us to better identify the differences between the two indexing databases, thus viewing digitalization and sustainable economy from multiple angles. Also, the combined analysis performed in Python worked with filtered data (the duplicates were eliminated) and brought new analyses, like N-Gram, Treemap, or AI Analysis.

3.2. Data Collection

Similar searches were carried out in the two databases that were used for indexing. In WoS, the title and the abstract and the author keywords matched the search (“digitalization” OR “digital*” OR (“artificial” AND “intelligence”) OR “AI” OR “blockchain” OR “cloud”) AND (“sustainability” OR “sustainable” OR (“triple” AND “bottom” AND “line”)). Additional filters were put on the document’s type (Article) and the languages (English). On 21 August 2024, when the query was performed, 1461 documents that were published between 2004 and 2024 were found. Because 97.60% of the documents were published between 2017 and 2024, this interval was chosen as the reference one; therefore, 1426 documents were selected to be analyzed. The year 2017 brings an increase of 1500% in terms of the number of records compared to 2004.

In Scopus, exactly the same search, with the same filters, produced 1939 results on August 21, 2024. The papers were published since 1996 but became relevant from 2017. In total, 97.22% of the documents were published between 2017 and 2024, the same chosen interval as for WoS; therefore, 1885 documents were selected to be analyzed. The year 2017 brings an increase of 1300% in terms of the number of records compared to 1996, for this Scopus query.

The search string that was used to retrieve relevant articles from WoS and Scopus includes also equivalent terms for digital, because in the literature, they can be considered synonyms or part of digitalization (AI, blockchain, cloud). The same thing is defined for sustainability or triple bottom line, as explained in [56].

When formulating the queries on both databases, we considered the two main topics of interest: digitalization and sustainability. For each term, synonyms were also included, as well as the family of words, in the case of digital*—with a wildcard, which could replace the following words: digital, digitally, digitalisation, digitalization, digitalized, digitalize, digitalism, digitization, digitize, digitized. Thus, we can exclude language ambiguity (digitalisation in British English or digitalization in American English) or technical slangs (triple bottom line for sustainability, as [56] uses it).

Data cleaning and pre-processing were performed in Python for the combined dataset. The first dataset (the Scopus document) was read from a .bib file. From the total 1885 records, 1867 were properly read (they did not have incomplete data). Data coming from the second dataset (the WoS document) were read and merged with the first one, resulting in a complete document of 2021 records. Overall, 154 new documents were added from the WoS database, where everything could be read correctly and no records were lost during the importing process.

3.3. Tools and Software

The applications available for bibliometric analyses in the field of research are limited in number; consequently, in this paper, the parallel analysis was carried out with the help of the “Bibliometrix” package from the R programming language. With the help of the Shiny application “Biblioshiny”, the multitude of commands that were already included in the “Bibliometrix” package have been integrated into an interactive web interface that facilitates the generation of visualization and the access to specific indicators [57]. This interface offers the user eight categories of possible analyses, divided according to the object of the research (sources, authors, documents) and the topic of the study (conceptual, intellectual, and social structure): Overview; Sources; Authors; Documents; Clustering; Conceptual Structure; Intellectual Structure; and Social Structure.

Each of these categories has one to twelve possible views, which can be easily adjusted based on the outcome that is intended. For most visualizations, configurations such as the number of evaluated records, the clustering technique, the time interval, and the graphic parameters can be changed to display the most important information in the most visually engaging way.

For the combined dataset, a bibliometric analysis was conducted using the pyBibX Python library, developed last year by a team of researchers from Universidade Federal Fluminense, a Brazilian university. The team consists of Valdecy Pereira, Marcio Pereira Basilio, and Carlos Henrique Tarjano Santos, who all presented a paper in 2023, in which they explain the capabilities of pyBibX and the benefits it brings to bibliometric and scientometric analysis [58]. They published the library on GitHub, for the first time, in April 2024 (release 3.3.1.). In June, Valdecy published the release 3.3.2, and on July 11, the newest release appeared (3.3.4).

pyBibX offers four main capabilities, which are all used in our analysis of the complete dataset coming from the two sources:

• correction and manipulation through operations like filtering (by the year, sources, countries, languages, and/or abstracts), by merging fields (authors, institutions, countries, languages, and/or sources that have multiple entries), or by merging different or the same database files one at a time;
• general analysis (like EDA (Exploratory Data Analysis) report; Word Cloud from the abstracts, titles, authors keywords or keywords plus; N-Gram bar plot from the abstracts, titles, authors keywords or keywords plus; Sankey Diagram with any combination of the following keys: authors, countries, institutions, journals, authors keywords, keywords plus, and/or languages; Treemap from the authors, countries, institutions, journals, authors keywords, or keywords plus, etc.);
• network analysis (Citation analysis, Collaboration analysis, Similarity analysis, World Map collaboration analysis);
• Artificial Intelligence analysis (like Topic Modeling using BERTopic to cluster documents by topic, visualize topics distribution, visualize topics by the most representative words, visualize documents projection and clusterization by topic, visualize topics heatmap, etc.).

As the authors mention in “pyBibX--A Python Library for Bibliometric and Scientometric Analysis Powered with Artificial Intelligence Tools” [58], the library comes with new capabilities for bibliometric analysis, which cannot be found in other tools like Bibliometrix, VOSviewer, SciMat, Scientopy, CiteSpace, Tethne, or many other specific packages. These unique analyses are N-Gram, Treemap, and the AI feature. In the opinion of the authors, the next best tool is Bibliometrix, by offering 11 out of 17 of the most important bibliometric key features.

3.4. Bibliometric Indicators

In Python, after we have obtained the complete and cleaned dataset, merged from Scopus and WoS, we performed the EDA report using the command report = bibfile.eda_bib().

word_cloud_plot method was used to derive a Word Cloud of 500 most important words from the abstract: bibfile.word_cloud_plot(entry = ‘abs’, size_x = 15, size_y = 10, wordsn = 300). The size of each word would reflect its frequency, thus being useful to quickly identify the most representative words, but which should be put into context. A check table was also generated in order to indicate the word and its importance in terms of frequency value.

The N-Grams plot is used for dividing a text into sequences of a predefined number of words in order to determine patterns. We have used the command bibfile.get_top_ngrams(view = ‘notebook’, entry = ‘title’, ngrams = 3, stop_words = [], rmv_custom_words = [], wordsn = 15), which extracts 15 combinations of base words, which contain 3 words that best identify the title.

The last plot that we used for the descriptive analysis on the complete dataset is the Tree Map. It can be performed on the keywords, authors, journals, countries, institutions, or keywords. With the command bibfile.tree_map(entry = ‘jou’, topn = 20, size_x = 30, size_y = 10, txt_font_size = 12), we generated a Tree Map of the 20 journals with the highest frequency.

In pyBibX, the network analysis focuses on interactions between citations, journals, countries, or authors. The adjacency analysis studies the interaction between such entities, and we used citation analysis to identify the citations between papers. In the network that is represented in the plot, the blue nodes represent documents, and the red nodes represent the citations. We used the command bibfile.find_nodes_dir(view = ‘notebook’, article_ids = [1880], ref_ids = []) to identify the paper with the id 1880’s citations and the command bibfile.find_nodes_dir(view = ‘notebook’, article_ids = [], ref_ids = [‘r_2629’]) to identify by which papers the document r_2629 is cited.

Collaboration analysis is represented in pyBibX as an interactive plot between authors, countries, and institutions. By using the code bibfile.network_adj(view = ‘notebook’, adj_type = ‘aut’, min_count = 8, node_labels = True, label_type = ‘name’, centrality = None), we obtained a collaboration analysis for the authors that interacted at least 8 times.

Similarity analysis (also represented as an interactive plot) can be performed using coupling or co-citation methods. The command that we used will perform a co-citation analysis to find documents that share common references. It finds documents that have at least ten references in common, helping to identify highly related documents based on both the topics they address and the sources they reference:

bibfile.network_sim(view = ‘notebook’, sim_type = ‘cocit’, node_size = 10, node_labels = True, cut_coup = 0.3, cut_cocit = 10).

World map collaboration analysis is used to visualize the patterns of collaboration between authors or institutions representing different countries. In the example, we show the countries with which the American researchers collaborated:

bibfile.network_adj_map(view = ‘notebook’, connections = False, country_lst = [‘United States of America’]).

The AI capabilities of pyBibX are especially offered by Natural Language Processing (NLP) algorithms. Thus, using the create_embeddings method, we created a list of 30 hot topics, based on the abstracts.

In the Biblioshiny interface, the graphs used in the analysis were generated with the help of the interactive menu that displays the eight analysis categories (Overview; Sources; Authors; Documents; Clustering; Conceptual Structure; Intellectual Structure; and Social Structure). Among these categories, six have at least one relevant graph that was selected for the research in order to complete the analysis’ overall perspective.

The “Overview” category graphs, which showcase the database’s general details, have been converted into tables to enhance the effectiveness of information extraction. This section’s indicators concentrate on the main information about the data, the authors, the annual production of papers, and the annual average of citations.

In relation to the database sources, the analyses were conducted using tables pertaining to the most relevant sources based on the number of publications and plots that take into account the sources’ production and their impact as seen through the perspective of the H-index.

The most relevant authors are presented in the section with the same name, according to the number of articles they published, their scientific production during the explored period, and the countries that have the most relevant corresponding authors.

The category related to documents encapsulated two tables presenting the 10 most cited articles from each database, thus providing important information on the research directions approached by the authors so far in this subject.

Using the keywords of the article as a framework, the “Thematic map” plot from the section “Conceptual Structure” has been selected to express the status and relevance of the themes for research. The configurations of this plot remain the default ones, with the “Keywords Plus” analysis field and the “Walktrap” clustering algorithm.

For the “Social Structure” section, both available plots were included in the analysis. Collaborative networking between countries and the collaborative world map present the same information in different visual forms. These two graphs show the way in which the authors of different countries collaborated with each other during the selected period. For “Collaborative networking between countries”, we kept the default settings, except the clustering algorithm parameter, which, in this case, becomes “SpinGlass”.

These graphs and tables generated by Biblioshiny, along with the ones obtained with pyBibX, address the essential details required to answer the research questions.

4. Results

This study reveals important trends in digitalization research connected to sustainability, based on data from 2017 to the present. Analysis of the combined WoS and Scopus dataset highlights major themes such as “sustainable development goal” and “the impact” as central to the discourse. Platform-specific analyses further uncover unique trends, with WoS having as major themes “dynamic capabilities”, “knowledge”, and “information-technology”, while Scopus highlights “China”, “human”, and “article”. These findings underscore the varied focus areas and research trajectories across platforms, contributing to a deeper understanding of the field.

4.1. Descriptive Analysis

Our descriptive analysis performed on the whole dataset consists of the EDA report; the Word Cloud from the abstracts, title, and keywords; the N-Grams plot; and the Tree Map.

In

Table 1. EDA report.

Main Information	Results	Main Information	Results
Timespan	2017–2024	Average Documents per Year	252.62
Total Number of Countries	111	Total Number of Authors	6527
Total Number of Institutions	3397	Total Number of Authors Keywords	5523
Total Number of Sources	775	Total Number of Authors Keywords Plus	5706
Total Number of References	7908	Total Single-Authored Documents	187
Total Number of Documents	2021	Total Multi-Authored Documents	1834
--Article	1998	Average Collaboration Index	3.64
--Article in Press	1	Max H-Index	8
--Article; Book Chapter	19	Total Number of Citations	42,050
--Proceedings Paper	3	Average Citations per Author	6.44
Average Documents per Author	1.13	Average Citations per Institution	12.38
Average Documents per Institution	2.28	Average Citations per Document	20.81
Average Documents per Source	2.58	Average Citations per Source	54.2

, it can be seen that from 2021 documents, we have 1998 articles, 1 article in press, 19 book chapters, and three proceeding papers, even if the filters in the indexing databases were put only on articles. It means they were allocated with several tags, which included article. From this report, we can conclude that each author had on average 6.44 citations, and each document was cited on average almost 21 times. A total of 111 countries were involved in the publication of articles with the topic of digitalization and economic sustainability in the years between 2017 and 2024.

The Word Cloud in Figure 2 represents the 300 most important words (or base words) for the analyzed subject. SUSTAIN appears with the coefficient one, which means each article had this word in its abstract. The word “digit” appears with a coefficient of 0.8764, while “develop” is in 42.055% of the abstracts. In the top 5, we also have “studi” with a coefficient of 0.03851 and “technolog” with a coefficient of 0.36.

Before we applied Word Cloud and N-Grams, we used the nltk package with the WordNetLemmatizer and PorterStemmer modules for lemmatizing and stemming the dataset. Lemmatization is good for treating similar words as one, based on their dictionary form (lemma), on their context, and on their part of speech. Stemming involves cutting off prefixes or suffixes from words to reduce them to their base form. For instance, digital and digitalization were transformed into digit and sustainability and sustainable into sustain.

We applied this technique to abstract, title, author_keywords, and keywords columns, and we did not extend it to many analyses because either the meaning is important to be kept or processing the text data in advance is not relevant. Also, misleading base words can be generated. For instance, we encountered a “thi” base word with a large frequency, and since we could not establish a relevant correlation with the original words, we had to manually eliminate it.

In Figure 3, we represent a 3-Gram plot for the title, in which we find the three base words that are the most associated in titles, and we have the-role-of with a frequency of 95; sustain-develop-goal with a frequency of 84; and the-impact-of, which appears in 68 articles.

Figure 4 presents the Tree Map for the 20 most important journals in the field of digitalization and economic sustainability. With 453 articles published in the last 7 years, Sustainability is the obvious leader. At a far distance, we find The Journal of Cleaner Production (with 54 published articles), and Technological Forecasting and Social Change (with 40 published articles).

4.2. Network Analysis

For the network analysis, we used in Python the Highlight citation analysis between documents, the Collaboration analysis, the Similarity analysis, and the World map collaboration analysis, as plots obtained with pyBibX methods.

4.2.1. Citation Analysis

For the citation analysis, we present two plots in Figure 5 and Figure 6. First, we identify the 13 citations in the document 1880 [59] and then the 10 documents in which the citation 2629 [60] is used.

4.2.2. Collaboration Analysis

The collaboration analysis is represented in detail in Figure 7. Twenty six collaboration clusters were obtained, and cluster 2 can be seen in Figure 7. Kozlova Valerija and Leal Filho Water have the largest number of collaborators (22 each).

A similarity analysis was also performed using the co-citation method, showing distinct clusters and their inner and inter collaboration.

4.2.3. World Map Collaboration Analysis

In Figure 8, we can see that the most productive country, in terms of published papers on the topic digitalization and sustainability, is China, with over 50 papers, followed by India, with around 40 papers. In this example, we analyze the collaboration between the USA and the other countries, and we can identify such countries like Brazil, India, or Spain.

4.3. Artificial Intelligence Capabilities

In Python, we have used the nltk package with the WordNetLemmatizer and PorterStemmer modules for lemmatizing and stemming the dataset before applying the NLP algorithm for clustering the topics.

AI capabilities were used to generate 30 interest topics, based on the available abstracts. In Figure 9 and Figure 10, it can be seen that the first topic (−1) is representative for 718 papers, more than 35% of the papers. It includes the words digit, sustain, thi, develop, studi, and research.

Figure 9 contains the graphical representation of the topics with a few of the representative words, while Figure 10 has a graphical topics projection, by showing how much a topic is representative for the domain in comparison with the others. We can see that the first topic (−1) and topic 0 have a huge influence on the researched field.

4.4. Comparison Between Web of Science and Scopus

Table 2. Main information about data: (a) WoS, (b) Scopus.

(a)
Indicator	Value
Timespan	2017:2024
Sources	480
Documents	1426
Annual Growth Rate	57.93%
Average Citations Per Document	17.56
Author’s Keywords	4425
(b)
Indicator	Value
Timespan	2017:2024
Sources	714
Documents	1867
Annual Growth Rate	69.08%
Average Citations Per Document	21.66
Author’s Keywords	5223

and

Table 3. Main information about authors: (a) WoS, (b) Scopus.

(a)
Indicator	Value
Authors	3974
Authors of single-authored docs	116
Authors of multi-authored docs	3858
Authors per document	3.45
International co-authorships (%)	38.64
(b)
Indicator	Value
Authors	5408
Authors of single-authored docs	158
Authors of multi-authored docs	5250
Authors per document	3.67
International co-authorships (%)	38.99

provide an overall perspective on the research level in this particular domain. Table 2 presents an overview of the data, with the element of interest in the topic being supported by the high values of indicators like “ Average citations per document”, “Documents”, and “Sources.” This premise is reinforced in both databases by the high values of the “Annual Growth Rate” indicator (57.93% and 69.08%), implying an increasing enthusiasm towards researching the correlation between sustainability and digitalization. The growth is considerably noticeable in the Scopus database, which may indicate the authors’ inclination to index in this particular database.

Table 3 contains the most important information about the authors, highlighting general information such as the number of authors integrated in each database but also the predilection of the authors towards collaboration or publication of Single-Country articles. Therefore, (a) provides a representation of papers from the Web of Science database, with an emphasis on the increased number of authors (3974) and the tendency towards international collaboration, since cooperation of this kind was identified in a proportion of 38.64% in the extracted database. Although the Scopus database has a larger number of authors (5408), it shows the same trends as WoS: the majority of authors published in the 2017–2024 interval multi-authored articles (5250), while the number of authors per document and international co-authorship values are slightly higher due to the larger number of researchers analyzed. When comparing the number of authors of single-authored papers (116) against multi-authored papers (3858), there is a significant difference that suggests likewise a propensity towards collaboration on the research of the area.

Although the number of papers published increased from 2017 to 2024, the average number of citations in each year shows a more complex shape (

Table 4. Annual scientific production and average citations: (a) WoS, (b) Scopus.

(a)
Year	Average Citations per Year	Number of Articles
2017	3.9	16
2018	5.7	23
2019	11	43
2020	11.8	97
2021	7.6	166
2022	6.7	279
2023	4.6	410
2024	1.6	392
(b)
Year	Average Citations per Year	Number of Articles
2017	5.6	14
2018	9.3	32
2019	13.3	54
2020	13.6	135
2021	10.8	199
2022	8	333
2023	6	547
2024	2	553

). The growing trend, with the difference from year to year becoming increasingly considerable, highlights the continuous growth in the publishing of papers. With 410 papers published, 2023 is the year with the most articles published in the WoS database, although Scopus has a perfect increasing trend, reaching the highest number of articles in 2024 (553)—and counting. In the context of a subsequent bibliometric investigation, it is expected that the year 2024 will maintain the existing trend in WoS as well, depending on the time of year with the highest number of publications. In contrast, the discussion is rather different when it comes to the annual average citations. The trend of average citations is increasing until 2020, when the maximum value recorded over the analyzed time frame is achieved (11.8 for WoS and 13.6 for Scopus). Later, following the year 2020, there is an abrupt decrease in the number of citations, which has lasted until the present.

4.4.1. Sources

The difference between the main sources found in this database can be seen in

Table 5. Top 10 most relevant sources: (a) WoS, (b) Scopus.

(a)
Source	Number of Articles
Sustainability	370
Journal of Cleaner Production	53
Technological Forecasting and Social Change	32
Business Strategy and The Environment	28
Sustainable Development	20
Heliyon	18
Sustainable Development of Modern Digital Economy: Perspectives from Russian Experiences	15
Annals of Operations Research	14
Resources Policy	14
Environmental Science and Pollution Research	12
(b)
Source	Number of Articles
Sustainability (Switzerland)	436
Journal of Cleaner Production	54
Technological Forecasting and Social Change	39
Business Strategy and The Environment	31
Resources Policy	27
Sustainable Development	23
Heliyon	19
Sustainable Cities and Society	18
IEEE Transactions on Engineering Management	16
Technology in Society	15

. In the WoS database, with a difference of 317 published documents, Sustainability is the most relevant source (370 documents), followed by Journal of Cleaner Production (53 documents), Technological Forecasting and Social Change (32 documents), and Business Strategy and The Environment (28 documents). The majority of the articles in the Scopus database are also from Sustainability, with 436 out of 1867 analyzed. It is worth noting that in both cases, the most significant difference is found between the journals in the first two positions, the following being close in terms of the number of published articles.

The distinction between Sustainability and the other journals is also noticeable in the number of papers published in the reference period by these relevant sources to the study. This is supported by an abrupt rise beginning in 2019 and a rapidly increasing trend following (Figure 11). The trend for the other four sources remains consistent until 2022, when it starts to modestly increase, particularly for the journal Sustainable Development (for WoS) and the journal Resources Policy (for Scopus), where the upward trajectory is more noticeable. Similarities, such as patterns in article production, indicate the authors’ overall tendencies, independent of the database in which the articles appear. As a result, two reliable sources indicate that interest in this subject matter and in the journal Sustainability began to increase after 2019.

In terms of the Hirsch index, Sustainability has the greatest impact recorded in the Web of Science database between 2017 and 2024, with an index value of 37, followed by Journal of Cleaner Production (24) and Business Strategy and The Environment (16). Similarly, Sustainability (46) and Journal of Cleaner Production (23) rank first and second, respectively, in the impact of the sources in the Scopus database, with a more noticeable difference this time, followed by the journal Technological Forecasting and Social Change (Figure 12). Considering that the H-index measures productivity by counting the number of publications and citation records, this aspect is essential in the analyses.

4.4.2. Authors

An overall assessment of the level of research on an area of study may be reinforced through studying information on authors. Therefore, bibliometric studies can benefit from key elements like the preferred cooperation techniques, the nations that have contributed the most, or the authors that have had the most influence.

The authors who have made the most significant contributions to the publication of articles, particularly about the implications of digitalization on sustainability, are highlighted in Figure 13. With 19 (WoS) and 21 (Scopus) articles published in each database, Li Y is the author with the largest contribution in both instances. Although the top 10 authors are very similar in the databases, there are no longer significant distinctions between the hierarchical positions in Scopus.

The most relevant authors are also identified through the Hirsch index in Figure 14. The H-index can rank authors in a relevant way based on both the number of published articles and the number of citations acquired over the time period under consideration, making it relevant to the analysis. The author with the greatest influence is consistent between the two databases: Gupta Shivam has an index of 8 in Scopus and 6 in WoS. The difference between the first places varies depending on the source. For WoS, the first three authors (Gupta Shivam, Kineber Ahmed Faraouk, and Oke Ayodeji Emmanuel) in the ranking have the same index (6), followed by Gunasekaran Angappa and Luthra Sunel, both with a score of 5. In the case of Scopus, the top position is claimed by a single author, followed by the second position, where Gunasekaran Angappa has an index of 7, and the third position, where a total of five authors with a score of 6 are identified. This research reveals an important fact: while the majority of the authors appeared in both graphical representations, the impact recorded in each database differs; the impact recorded at the Scopus level is more significant than that in WoS.

Figure 15 assesses 2023 as the most successful year in terms of production across the entire time frame, considering that the ten most relevant authors produced the largest number of papers and received the highest number of citations. The maximum point of published articles and citations is shown by the largest dot on the first line of the graph, which has the most intense color. As an acknowledged author in the area, Li Y gained recognition for her work on sustainability and digitalization through seven articles published in 2023 in WoS and eight in 2023 in Scopus, with a total of 322.5 citations (155 in WoS and 167.5 in Scopus). Kumar A and Liu Y topped the number of articles published in Scopus in 2023, each with nine articles. The other authors usually produced two papers annually, independent of the source database. The presence of distinct graphic tools with consistency in the second section of the graph indicates a collective interest in an investigation of the two transitional processes. As a result, after 2022, all authors from the top 10 published articles in both databases, regardless of citation count.

Bibliometric analyses can benefit significantly from knowledge of the authors’ countries, since they highlight the regions with the greatest influence and collaboration preferences, as well as the global interest of the area under study. In this instance, researchers from China have made a significant contribution, accounting for almost 25% of all the papers in the WoS database. Single-Country Publications (SCPs) articles are more common in China than Multi-Country Publications (MCPs) ones (

Table 6. Top 10 most relevant corresponding authors’ countries: (a) WoS, (b) Scopus.

(a)
Country	Articles	SCP	MCP	Percentage of Total
China	357	245	112	25.04%
Italy	84	59	25	5.89%
India	70	36	34	4.91%
Germany	61	40	21	4.28%
United Kingdom	61	19	42	4.28%
Spain	56	38	18	3.93%
USA	53	31	22	3.72%
Korea	40	29	11	2.81%
Romania	35	30	5	2.45%
Australia	33	18	15	2.31%
(b)
Country	Articles	SCP	MCP	Percentage of Total
China	419	275	144	22.44%
Italy	104	72	32	5.57%
India	101	62	39	4.41%
United Kingdom	78	28	50	4.18%
USA	73	43	30	3.91%
Germany	63	43	20	3.37%
Spain	62	39	23	3.32%
Korea	58	37	21	3.11%
Australia	37	20	17	1.98%
Saudi Arabia	37	16	21	1.98%

). Overall, the preference of the top 10 countries is towards SCP, the exception being the UK, where out of the 61 published articles, 42 are of the MCP type, while only 19 are SCP. Similar results were observed in the case of Scopus, where China is likewise the country with the largest contribution, capturing only 22.44% of all published articles. Additionally, the trend of publishing SCP articles persists, with Saudi Arabia and the UK being the outliers. Compared to the other countries in the top 10, these countries are more likely to have solid international connections. Similar results were obtained by other similar bibliometric papers, such as that of Zhou et al. [61]. In the previously mentioned analysis, the results surprise, as China is the country that placed first in the production of articles, followed by United Kingdom. An interesting finding of the study is that, although being in the top two, China and the United Kingdom do not have close scientific collaborative relationships with the rest of the countries. This finding contradicts the large number of MCP papers from China and UK that were published between 2017 and 2024 in the current research. Although China looks to be more likely to provide SCP papers, the number of MCP articles remains significant (112 WoS and 144 Scopus). These disparities might be attributed to the time frame studied, since Zhou’s research ended in 2021, while in our analysis, a rapid increase was noticed in 2020, which was sustained until 2024, a prosperous period, during which these countries’ international relations expanded.

International collaboration can result in enhanced research opportunities and the proposal of innovative research solutions, drawing on factors from various economic and social branches. These factors play a decisive role in ensuring harmonious and fruitful collaborations.

A common challenge in establishing international relations is accessing funds. Financial support is very relevant in this case and can be received from various sources such as government organizations that aim to increase international relations with certain countries, funds from international companies, or private investments.

Cultural ties play an important role also, as authors often choose to collaborate based on the proximity of the countries and cultural similarities. For example, collaborations between countries that have the same national language are more common than in cases where communication must be carried out exclusively in an international traffic language such as English, as coordination between authors can be difficult.

Technological progress and digital infrastructure can have a strong impact on establishing international collaboration. In addition, the economic context is a very relevant factor, as authors from countries with a stable economic context free of tensions can benefit from collaboration opportunities that can strengthen diplomatic relations between countries [62].

Having previously stated that China, Italy, and India are the top three most productive countries in terms of publications in the fields of digitalization and sustainability, Figure 16 and Figure 17 provide more information about the countries’ collaborations. Figure 16 confirms that China not only has the highest number of publications (dark blue color), but it also maintains solid collaborative connections with other areas, as seen by the numerous collaborations highlighted with red lines. Figure 17 paints a substantially more complex picture of global collaboration by clustering all the countries included in the analysis. Four clusters are therefore created based on collaborations that have been colored differently. China comes out on top once more, particularly when it comes to connections with the US, UK, India, and Pakistan. Despite having a large number of articles, Italy, the country in the second position, does not have as many close connections as China. It collaborates mainly with the countries that are part of the same cluster, marked in green, and the closest connection is found in the relationship with France. Although it ranks fifth in terms of article production, the United Kingdom competes with China in terms of international collaboration, through significant connections to China, India, and France. The findings are consistent with other papers in the specialized literature that cover similar subjects. Lozano-Raminrez et al. [63] pursued a bibliometric analysis of the Scopus database, and the results at the level of countries with significant contributions show China leading with a percentage of 15.6%. Differences in ranking are visible since the hierarchy in this particular scenario is China, USA, UK, Italy, and Australia. We can attribute the disparities in percentages and rankings to the restriction of the terms employed in database extraction. Similarly, Irajifar et al. [47] examined bibliometric data from papers on digitalization and sustainability in the WoS database. With greater variances, the ranking at the country level is overall different: USA, China, and UK, and this variation is not only due to the terms employed, but also to the paper’s far more extensive time span (1900–2021).

Considering that both countries are outstanding in distinct areas—publication of papers and strong collaborations—it is not surprising that China and the UK have the most productive collaboration, having published 30 papers together. The next groupings with a strong connection in terms of the number of published articles are China–Pakistan with 29 articles, China–USA with 29 articles, and India–UK with 28 articles, followed by collaborations under 20 articles. It is relevant to mention that over 40 collaborations between countries have a single published article; these connections potentially reflect the possibility for more than 40 successful collaborations that might raise the level of research. As a result, the current state of research might be enhanced by a significant number of MCP publications.

Due to the near identical links formed between the nations under analysis, it is difficult to overlook the map-level parallels between WoS and Scopus. This is to be expected, as publication databases and journals often have minimal impact on the preferences of authors exploring international collaboration. As a result, Figure 18 appears very similar to Figure 16, illustrating the numerous collaborations related to China.

The resulting collaboration network separates the investigated countries into four clusters, with China standing out as the most dominant. There are a few notable outliers, but generally speaking, different countries only collaborate with other members of their same cluster. The most substantial collaborations in Figure 19 are those that are formed between clusters; they include collaborations between China and other countries (UK, India, USA, and Pakistan), as well as connections between the UK and India.

In the combined dataset (Figure 8), we identified China as the most productive country in terms of published papers on digitalization and sustainability, with over 50 publications, followed by India with approximately 40 papers. This analysis was also supported by the parallel analyses of the Scopus and WoS databases. The focus on China was also maintained in the two collaboration networks between countries, which identified strong partnerships with UK, Pakistan, USA, and India, regardless of the database.

4.4.3. Keywords

The parallel analysis in Bibliometrix for the thematic maps will be used to answer the following hypotheses:

H1: 

Motor themes are more prevalent in both WoS and Scopus databases than niche themes.

H2: 

Keywords related to sustainability have higher density in the Scopus thematic map than in WoS.

H3: 

There is a parallel evolution in the research themes across WoS and Scopus, with both databases showing an increasing emphasis on ‘digitalization’ and ‘sustainable development’ over time, but with varying levels of centrality and density.

The thematic map of WoS is displayed in Figure 20, taking into account the Keywords Plus used in the database. The map, defined by internal (density) and exterior (centrality) associations, is divided into four quadrants, as follows [64]:

• Niche Themes: high density and low centrality
• Emerging or Declining Themes: low density and low centrality
• Motor Themes: high density and high centrality
• Basic Themes: low density and high centrality

The density of the map assesses the level of theme development using internal associations of keywords, and centrality also measures the relevance of the themes, but based on external associations of keywords [65]. The motor themes are considered in the specialized literature to be mainstream, excelling in both directions described by this type of graph. Considering the keywords plus, the motor themes are “dynamic capabilities”, “knowledge”, and “information-technology”, followed shortly after by “supply chain management” and “firm performance”. At the opposite pole, declining themes exhibit low values for both density and centrality, the specific themes being “growth”, “ict”, and “economic-growth”. The placement of these three terms in the graph highlights the research potential of these areas that have yet to achieve their full potential. A distinguishing feature of the study stands out regarding the other two more balanced quadrants: two theme clusters were identified for each of them. For the niche themes, the first cluster contains two keywords, “co2 emissions” and “urbanisation”, and the second cluster focuses mainly on the social aspects of the research with the keywords “social media”, “behavior”, and “user acceptance”. Finally, the themes regarding subjects like management, internet, and design are seen as basic, as the suggestive name of this quadrant accurately describes the commonly found nature of the articles that address these issues. The cluster with the largest number of keywords is marked with orange in the basic themes cluster, with the number of appearances of the first term (“management”) reaching 216 articles and “impact” reaching 157 articles. In contrast, the smallest cluster is located in the niche themes quadrant, with only two terms appearing in 21 articles in 2017–2024, both of them focusing on the sustainable aspects of research.

Regarding the Scopus database, the differences in the thematic map are visible in Figure 21. This time, for each quadrant, there is only one cluster, and the size variations are significant. The niche themes encapsulate “energy efficiency”, “energy utilization”, and “smart city”. Therefore, a narrow group specialized in this field investigates segments that focus on aspects related to sustainable energy use and sustainable development of cities. Unexpectedly, the keywords included in the declining themes include “sustainability”, “artificial intelligence”, and “digitization”. In addition to the keywords within these themes, it can be noticed that the declining themes and the basic themes share the same scientific pool, since the basic themes include comparable subjects such as “sustainable development”. The terms “China”, “human”, and “article” are particularly prevalent in the subjects that are being adopted more frequently in publications. The term “China” is a relevant aspect regarding the thematic map given the fact that the authors with the most significant contribution in the area are from China. As a result, a big proportion of the studies could be focused on China’s context, thus having the keyword be the name of the country. By having the term “article” as a keyword, the studies highlight the source of the data in the academic literature, for instance, having as a main object the article when extracting the database for systemic reviews. In the cases when the research highlights “human” as a keyword, the contents often take into account the behavior or the implications relevant to human populations in relation to processes such as digitalization and sustainable transition.

Both thematic maps helped us to test and answer the proposed hypothesis; thus, for H1, Figure 20 (WoS) confirms that “dynamic capabilities” and “information technology” appear in the motor themes quadrant with high density and centrality, supporting the hypothesis. In contrast, niche themes like “co2 emissions” and “urbanization” are placed in the niche themes quadrant, with low centrality and high density. For Scopus, as well, “china”, “article”, and “human” are dominant motor themes, compared to the niche ones: “energy efficiency” or “smart city”.

High-density keywords appear in the niche and motor themes. In WoS, we find “co2 emissions”, “social media”, or “dynamic capabilities”, while in Scopus, there are more sustainability-related keywords, like “energy efficiency” and “smart cities”. H2 is supported by these results.

H3 is also supported because, in WoS, digitalization appears in the motor themes quadrant as “information technology”, indicating a significant impact in the field of research and sustainable development is merely represented by “economic growth” as an emerging theme with relatively high centrality. In Scopus, the term “digitalization” is found in the emerging themes quadrant with relatively lower centrality compared to WoS, while “sustainable development” is positioned in the basic themes quadrant with high centrality but lower density, implying that it is an important topic.

4.5. Most Cited Documents

This section will examine the literature on the implications of digitalization in the development of sustainability, from the perspective of the 10 most cited articles.

4.5.1. Web of Science Literature Exploration

Table 7. Top 10 most cited articles: WoS.

Paper (First Author, Year, Journal, Reference)	Number of Authors	Region	Total Citations (TC)	Total Citations per Year	Normalized TC (NTC)
Saberi, Sara, 2019, International Journal of Production Research [67]	4	USA	1523	253.83	23.05
Kouhizadeh, Mahtab, 2021, International Journal of Production Economics [68]	3	USA	535	133.75	17.58
Esmaeilian, Behzad, 2020, Resources, Conservation and Recycling [69]	4	USA	355	71.00	6.03
Bag, Surajit, 2021, Technological Forecasting and Social Change [70]	4	South Africa, France, UK	332	83.00	10.91
Bai, Chunguang, 2020, International Journal of Production Research [71]	2	China, Finland	331	66.20	5.62
Luo, Shiyue, 2023, Business Strategy and the Environment [28]	6	China	270	135.00	29.13
Upadhyay, Arvind, 2021, Journal of Cleaner Production [72]	4	UK, Turkey	268	67.00	8.80
Nishant, Rohit, 2020, International Journal of Information Management [29]	3	Canada	262	52.40	4.45
Leng, Jiewu, 2020, Renewable and Sustainable Energy Reviews [73]	7	China	249	49.80	4.23
Di Vaio, Assunta, 2020, International Journal of Information Management [74]	2	Italy	232	46.40	3.94

captures the most relevant details regarding the most cited documents extracted from this database on the implications of digitalization in the transition to sustainability, including information about the authors, the total number of citations (TC), the total number of citations per year, and normalized total citations (NTC). NTC refers to the total number of citations attributed to each of the authors of an article, assuming that the effort of all authors is equal, while considering the average citations per document for every article published in the same year [66]. To understand the evolution of the implications of digitalization in sustainability, a review of the top 10 most cited documents globally has been carried out, together with the direction implemented in each paper with the selected keywords.

In 2019, Saberi published the most cited article [67] recorded in the Web of Science database between 2017 and 2024. This article addresses the issue of globalization in terms of supply chain management, as the complexity of this field is gradually increasing, becoming more difficult to control from a management point of view. The digitalization concept is present through the studied solutions: blockchain technology and smart contracts. The authors consider the potential barriers that could occur during the implementation of blockchain technology at the business and supply chain levels, considering the early stages of blockchain technology in 2019. Another important topic addressed by this article is the necessity of ensuring that the relevant sustainability requirements are met. As a result, societal pressure from the community and the government requires the blockchain implementation to have a low environmental impact. This article received, by the end of the reference period, 1532 citations, with an average of 253.83 citations per year and 23.05 normalized citations. All four authors of this paper are from the USA, indicating that this is a Single-Country publication.

The article placed second [68] in the ranking is based on citations addressing topics similar to the first one. Therefore, issues raised regarding the impact that the digital transformation of the supply chain has on the environment and the possible obstacles that can be encountered in the blockchain implementation in businesses are studied by various methods in this article as well. An important connection occurs between these two articles, as Kouhizadeh’s article follows Saberi’s article, by using the concepts analyzed initially. Consequently, Kouhizadeh incorporated the barriers identified by Saberi and developed a survey around them in order to find trustworthy solutions for the potential limitations described. The decision-making methodology used refers to the DEMATEL model, which is specific to supply chain problems. Numerous barriers were identified and later grouped into four specific categories that include: “Technological context”, “Organizational context”, “Environmental context (Supply chain view)”, and “Environmental Context (External view)” [39]. Following the application of the DAMATEL method, the paper generated six propositions that can facilitate decision making in the situations presented through these barriers. By the end of the reference period (2024), the paper collected 535 citations, with an average of 133.75 citations per year and normalized citations of 17.58. Similarly, this paper is a Single-Country publication, since the three authors are from the same country (USA).

Esmaeilian addressed in his paper [69] notions about Industry 4.0, Blockchain, and sustainability. Similar to the articles already mentioned, the interest of the researchers is the development of supply chain technologies in the direction of improving sustainability. This study aims to provide an overall summary of the effects of Industry 4.0 and blockchain on the expansion of the supply chain’s sustainability area. As a result, the article effectively formulates a series of recommendations for potential research directions in the subject by reviewing the level of studies in the field and emphasizing the areas that need more exploration. The article was written by four authors, all from the USA, and received 355 citations, an average of 71 citations per year, and a normalized total of 6.03.

Bag [70] believes that supply chain performance is closely linked to big data analytics-powered artificial intelligence, and based on the literature review, it emerged that the reasons why manufacturing organizations adopt these performance improvement methods are not sufficiently researched. The authors’ focus in this work is on the statistical validation of the hypotheses formulated for a set of South African manufacturing companies. The article offers new insights regarding the motivation for the adoption of big data analytics-powered artificial intelligence and their impact on manufacturing sustainability. The most relevant reason identified as the basis of these technological changes proved to be the pressure from external factors such as governmental ones for the development of Industry 4.0-specific skills at the company level. For this article, four authors collaborated, from South Africa, France, and the UK, indicating a Multi-Country publication. With similar results, the paper collected 332 citations in total, with an average of 83 citations per year and 10.91 normalized citations.

Bai and Sarkis [71] introduced in 2020 a new hybrid group decision method designed to evaluate and select the right blockchain technologies. The article considers that blockchain technologies can ensure supply chain transparency to support sustainability. To facilitate the understanding of blockchain technology selection decision concepts, the paper introduces an illustrative way of visualizing the sensitivity analysis. Through the perspectives delivered, the authors provide groundwork for researching ways to evaluate blockchain technologies in the supply chain. Two authors, from China and Finland, collaborated in order to achieve 331 citations in total, an average of 66.20 per year and 5.62 normalized citations.

Luo et al. [28] explored through various methods of data analysis the impact that the digitization of the economy can have on the sustainability of a country, specifically, on China. In the context of the development of sustainability at the global level, the way digitalization influences the economy can have both direct and indirect effects on the green policies adopted by countries. Using various data modeling techniques, the positive effects that the digital economy has on environmental innovations in China are proven. Six authors from China collaborated for this research, collecting 270 citations by the end of the reference time period, with an average of 135 citations per year and normalized total citations of 29.13.

Upadhyay et al. [72] presents blockchain as a revolutionary technology, which is able to update and share the information using databases in a decentralized manner, peer-to-peer, open-access network or using linking ledgers, ensuring the security of the data that are stored. Due to the rapid evolution of the technology, blockchain is now developed in numerous domains, and it became more and more important to understand its sustainability and the ethical implications of the blockchain implementation. The article focuses on identifying the blockchain positive impact, contributing also to the Industry 4.0 domain. The results showed an important impact for blockchain on the circular economy by significantly reducing the costs; increasing the performance of the supply chain; ensuring human rights protection; and, one of the most crucial benefits, reducing the CO₂ emissions. The four authors found that there are more benefits than challenges in the implementation of blockchain. At the end of the analyzed period, the article has a total of 268 citations, with an annual average of 66.20 citations and normalized citations of 5.62.

Nishant et al. [29] focus their research on AI, which has the potential to transform the industries and business practices, solving major problems in society, such as sustainability, since the degradation of the environment is more and more visible due to the climate crisis. Thanks to AI, the resources and energy of human activities can be reduced significantly. During the research, numerous challenges have been analyzed such as cybersecurity risks, adverse impacts of AI usage, and the response of humans to the AI. The results show that the implementation of AI for sustainability should take into consideration multilevel views, economic value, sociological considerations, and design, thinking that could help AI to provide solutions with immediate effect without affecting the environment in the long term. Three authors from Canada collaborated on this study, which reached 262 citations in total, with an average of 53.40 citations each year.

Leng et al. [73] examine the smart technologies, which could have a crucial impact on the manufacturing domain, keeping them sustainable. Blockchain was researched, while Industry 4.0 was defined from technical, operational, and organizational perspectives and is one of the recent technologies developed, which provides more sustainable business and industries processes. The purpose of the analysis is to explain how blockchain can offer the possibility of a sustainable manufacturing system and product lifecycle management. The paper analyzes the existing literature and the research in blockchain from a sustainable manufacturing perspective, but it is still at the beginning of the evolution, having a series of challenges from a technical perspective, which could lead to social barriers. Blockchain should demonstrate the benefits in industry and business, making them more sustainable. The largest number of authors registered in this ranking collaborated for this article, seven authors from China. The article received 249 citations in total, with an annual average of 49.80 and normalized total citations of 4.23.

Di Vaio and Varriale [74] explore the implications of blockchain technology in the operations management (OM) domain, facilitating the decision-making process in supply chain management (SCM), making it more sustainable. The airport industry includes both technologies, offering a sustainable performance. The Italian airport infrastructure was examined and adopted the Airport Collaborative Decision Making (A-CDM) system, which represents one of the main applications of blockchain technology in the airport industry, by assuring the collaboration between the main actors of the aviation domain and the air traffic controllers in order to reduce to minimum the inefficiency and the uncoordinated activities. The research also resents the existing concerns regarding OM and sustainability problems that exist in Italian airports. The two authors from Italy collaborated to write the article, which received 232 citations, with an annual average of 46.40 and normalized citations of 3.94.

4.5.2. Scopus Literature Exploration

Table 8. Top 10 most cited articles: Scopus.

Paper (First Author, Year, Journal, Reference)	Number of Authors	Region	Total Citations (TC)	Total Citations per Year	Normalized TC
Saberi, Sara, 2019, International Journal of Production Research [67]	4	USA	2046	341.00	25.69
Kouhizadeh, Mahtab, 2021, International Journal of Production Economics [68]	3	USA	671	167.75	15.54
Di Vaio, Assunta, 2020, Journal of Business Research [75]	4	Italy, Malaysia, France	483	96.60	7.12
Esmaeilian, Behzad, 2020, Resources, Conservation and Recycling [69]	4	USA	476	95.20	7.01
Bag, Surajit, 2021, Technological Forecasting and Social Change [70]	4	South Africa, France, UK	421	105.25	9.75
Bai, Chunguang, 2020, International Journal of Production Research [71]	2	China, Finland	405	81.00	5.97
Nishant, Rohit, 2020, International Journal of Information Management [29]	3	Canada	386	77.20	5.69
Singh, Saurabh, 2020, Sustainable Cities and Society [76]	6	South Korea, UK	357	71.40	5.26
Upadhyay, Arvind, 2021, Journal of Cleaner Production [72]	4	UK, Turkey	351	87.75	8.13
He, Bin, 2021, Advances in Manufacturing [77]	2	China	346	86.50	8.01

shows the 10 most cited articles from the Scopus database, including some of the most cited Web of Science articles from Table 7.

The most cited article in this database is, as in the WoS database, the article ”Blockchain technology and its relationships to sustainable supply chain management” [67] from 2019. In this case, 2046 citations were collected in total, with an average of 341 citations per year and normalized total citations of 25.69.

Also, the article ranked second coincides in the two databases. This was to be expected since these two articles ([67,68]) together provide a complete analysis of the barriers encountered in the implementation of blockchain in the supply chain and the ways to solve them. This article collected a total of 671 citations this time, since the analysis on the Scopus database extracts a larger number of citations overall. Notable figures include normalized total citations of 15.54 and an average of 167.75 citations annually.

Di Vaio [75] analyzes the impact of the Artificial Intelligence (AI) role in the construction of a sustainable business model (SBM), providing a quantitative overview of the academical publications on the domain, presenting also the collaboration among AI, Machine Learning (ML), and sustainable development (SD), having in mind how computer science can affect production and consumption patterns in order to obtain a sustainable resource management that fulfills the Sustainable Development Goals (SDGs) according to the United Nations (UN) 2030 Agenda. The database used during the research contains 73 publications in English, published between 1990 and 2019, and the results show that companies, universities, and governments should focus on the implementation of AI in SBM. There are four authors that collaborated to write the publication, from Italy, Malaysia, and France, having up to the moment of this analysis a total of 483 citations, with an average total citations per year of 96.60 and a normalized TC value of 7.12.

The article ranked third on the Web of Science list is referred to as being ranked fourth in the Scopus list. One noteworthy observation is that, despite the different ranking frameworks, the article with the fourth-place ranking has more citations (476) than the third-place item in the WoS database (355). The article [69] also manages to reach an annual average of 95.20 citations per year and normalized citations of 7.01.

Bag et al.’s article [70] is also previously mentioned in the Web of Science analysis, this time reaching a total number of citations of 421, a yearly average of 105.25, and normalized citations of 9.75. A similar situation is encountered in the case of Bai’s article [71], which manages to collect 405 total citations, normalized citations of 5.97, and an annual average of 81, and in the case of Nishant [29], who collected 386 citations in total during the reference period, there are normalized citations of 5.69 and an annual average of 77.20 citations, higher values than those already discussed in Table 7.

Singh [76] considers that smart cities can become intelligent by adapting emerging technologies, such as blockchain, which is rapidly evolving and could facilitate the shift to a new digital smart city ecosystem. Blockchain already can be applied to numerous solutions, such as Internet of Things (IoT), cryptocurrency, risk management, and financial services, while the combination of blockchain and AI could lead to a smart city built with sustainable ecosystems. The focus of the authors was to provide a comprehensive literature review of the blockchain domain, presenting the issues and problems that block the evolution of the system in smart cities. The key points that can be implemented in order to achieve an intelligent transportation system, by combining blockchain and AI, were discussed. In total, six authors collaborated during the research, from South Korea and the UK. The article has a total of 357 citations, with an average total citations per year of 71.40 and a normalized TC of 5.26.

The final item shared by both databases is Upadhyay et al.’s [72], which has 351 citations overall in the Scopus database, with an average total citations per year of 87.75 and 8.13 normalized citations.

He and Bai [77] explored intelligent manufacturing, which allows higher productivity, manufacturing flexibility, and quality with reduced costs, representing one of the topics that received significant attention in the last years. Digital twin is a technology implemented in intelligent manufacturing (IM), which captures it in real time. In order to fully understand the potential, the authors reviewed both technologies and explored the sustainability of IM, including also an application developed with IM based on the digital twin and presenting, also, the future research ideas for the IM. During the research, two authors from China worked, having a total citation of 346, with an average citations per year of 86.50 and a normalized TC value of 8.01.

By mapping the relationship between digitalization and sustainability, our research makes a significant contribution by providing a quantitative summary of the major trends, significant figures, and new subjects in the area. Nonetheless, the relevance of these findings is increased by a more comprehensive qualitative analysis. The prominence of particular authors or countries, for example, could be a reflection of larger institutional, cultural, or legislative frameworks that encourage innovation concerning environmentally friendly digital technology. Comparably, the most cited publications’ thematic focus identifies important areas of influence, such as the problems of globalization or artificial intelligence, which may be further investigated for their real-world applications and social ramifications. In addition to offering a descriptive map of the research landscape, the findings can offer practical insights into how digitalization can be used to accomplish targets for sustainability by relating these patterns to tangible issues like accomplishing the SDGs or addressing environmental issues such as CO₂ emissions.

5. Discussion

5.1. Contribution to Theoretical Research by Connecting Findings to RQ

The five research questions were addressed and answered by this research. Thus, RQ1 poses the question of distributing the publications and the citations over the period of time that was analyzed. In Table 4, it is presented that in 2023, the WoS database has the largest numbers of articles, while in Scopus, the most productive year is 2024. In terms of citations, 2020 is the year with the highest rate of citations, and it is totally explainable, considering that citations are made on older papers. Regarding the literature trends in this cross-cutting topic, previous studies also confirmed an exponential growth in the number of publications in the post-COVID period [47,48,78]. In this sense, Irajifar et al. [47] argued that digitalization has emerged as a means to address economic, social, and ecological crises, with accelerating climate change and environmental instability. In general, there was a diversification in the topics until 2019, when they shifted their focus to computer science, technology, and the environment. Initially, there has been scattered research on decision making, IT, and innovation, and the field experienced a leap forward after the introduction of SDGs in 2015, focusing on energy, extensive data modeling, and smart cities. Finally, the focus on new technologies using big data and industry 4.0 has rapidly grown in recent years.

RQ2 (What are the most influential articles, authors, journals, and countries in this field?) gets its answer in multiple analyses. Figure 13 identifies Liu Y and Kumar A as the authors with the highest number of published papers; Figure 8, Figure 17 and Figure 19 identify China as the most productive country, having published over 50 papers in the last seven years. Previous studies also pointed out the notable absolute advantage gained by China in recent years [47,48,78].

The Tree map in Figure 3 and Sources’ production over time in Figure 11 show that Sustainability is the main journal in which articles about digitalization and sustainability are published. Our findings are consistent with those of D’Adamo et al. [78], Irajifar et al. [47], and Davidescu et al. [32], where the journal Sustainability had the highest number of publications, followed by Journal of Cleaner Production. The top 10 most cited articles are presented in Table 7 and Table 8, and they both show that Saberi et al. [67] is the most cited paper, regardless of the indexing databases.

RQ3, regarding the collaboration trends, is satisfied by Figure 7, in which Gunasekaran Angappa (Pennsylvania State University) can be identified with high collaborations, while Figure 19 and Figure 17 show that China has the largest number of collaborations between countries, this being justified by the large numbers of papers published by Chinese authors. Increased cooperation between authors from two or more countries is also a trend in recent years, being in line with the current academic development rules of prestigious international journals.

The motor themes, which are considered to be mainstream, are quite different when analyzing the indexing databases. In WoS, the hot topics are dynamic capabilities, knowledge, and information technology, while in Scopus, they are China, article, and human. Thus, Figure 20 and Figure 21 solve RQ4 by also identifying the emerging themes—growth, ict, and economic growth for WoS and sustainability, artificial intelligence, and digitalization for Scopus.

RQ5 is solved by applying NLP in Figure 9, which identifies the best topic that would refer to the digit, sustain, thi, develop, studi, and research base words.

5.2. Contribution to Practice and Policy Implications

The widespread adoption of modern technologies generates new research directions, highlighting the importance of studying the role of digitalization in reaching sustainability. Our bibliometric analysis on digitalization and sustainability provides valuable policy implications for both policymakers and stakeholders in academia and industry.

On the one hand, research institutions and stakeholders in academia could identify and prioritize research trends based on our findings. The bibliometric analysis allowed us to highlight key research trends and emerging themes in the field of digitalization and sustainability during 2017–2024. It also underscores the need for interdisciplinary collaboration to achieve SDGs through digitalization. Academic institutions can use these insights to align research agendas with global sustainability priorities, budget research projects in niche areas, and develop digital tools to support sustainable practices.

On the other hand, these findings could also guide both governmental and corporate decision-makers to prioritize areas for development and further exploration, where digital technologies can accelerate progress toward SDGs. Policymakers could thus plan their sustainability strategies more effectively and design policies to sustain digital solutions, such as smart infrastructure, clean technologies, and data-driven governance. Governments can use these insights to allocate resources effectively, foster public–private–academic collaborations, and create regulatory frameworks that consider the trade-off between innovation and environmental and social responsibility. These digital technologies have the potential to drive the development of green business models while enhancing profitability through the digital transformation of production processes.

Moreover, our findings offer critical policy insights for stakeholders in industry by highlighting the potential for adopting digital tools, such as IoT, AI, and blockchain, to increase resource efficiency and reduce environmental impact. Industry leaders can leverage these findings to guide investment in sustainable digital solutions, prioritize expenses for research and development in areas with high sustainability impact, and adapt business models to meet regulatory and market demands for sustainability. Finally, the analysis emphasizes the need for partnerships between academia, industry, and policymakers to develop practical digital solutions that are in line with SDGs.

6. Conclusions

This paper addresses a critical knowledge gap by exploring the nexus between digitalization and sustainability, which gained increasing interest in both academic and business fields, especially since SDGs were first introduced through the 2030 Agenda for Sustainable Development. Unlike previous studies that mostly explored documents extracted from just a single platform, we conducted a more comprehensive and objective bibliometric analysis on the relationship between digitalization and sustainability by gathering information from two of the most scientifically well-known databases (WoS and Scopus) and performed the analysis on a combined and cleaned set of data. For comparison reasons with the previous literature, we also conducted a parallel analysis of the documents from each of the two databases. Moreover, we searched deeper into the policy implications of our research to argue how our bibliometric analysis findings could serve policymakers and stakeholders in academia and industry.

This study created a detailed map of the connection between digitalization and sustainability, emphasizing major trends, prominent themes, and geographic emphasis areas. Beyond these outcomes, it is critical to evaluate the qualitative effects. For example, the prominence of China-focused research highlights the country’s leadership in both digital and sustainability research while simultaneously raising concerns about regional discrepancies in outputs and international coordination. Similarly, the thematic emphasis on technologies such as AI and IoT illustrates their transformational potential, but it requires further investigation of societal concerns such as equal access to technology, data privacy, and the long-term social consequences of digitalization.

These findings not only align with the literature, confirming the reliability of the investigation, but also highlight how digitalization supports achieving all three dimensions of sustainability related to the SDGs. Following D’Adamo et al.’s [78] approach, we reckon that economically, digitalization aligns mostly with SDGs 7, 8, 9, 11, and 12, as it enables smarter resource management to support environmental sustainability (SDGs 7, 11, and 12); it drives economic growth through increased productivity, remote work facilities, and new entrepreneurship opportunities (SDG 8); and it promotes sustainable industrialization and fosters innovation (SDG 9).

From the social sustainability perspective related to SDGs, digitalization aligns mostly with SDGs 3, 4, 5, and 17. The implementation of digital platforms enhances access to education, healthcare, and information (SDGs 3, 4, and 5), empowering marginalized communities and fostering equity, while also facilitating global collaboration and partnerships (SDG 17) towards a sustainable future.

Finally, from the environmental sustainability perspective, digitalization has proven to align mostly with SDG 13, as it enables smarter resource management to support green economy through technologies like IoT and AI, which optimize energy use and reduce waste.

In conclusion, the results of our bibliometric analysis emphasize the critical role of digitalization in achieving SDGs and underscore the importance of governance and corporate decision-makers in these processes. However, the limited integration of SDGs into digital government transformation discourses suggests that this remains an area ripe for further exploration.

Although this study’s methodological approach provided significant results, it should be acknowledged that it has a number of limitations. First, the data extraction process was limited to the English language and particular terms, having as a source two specific databases, which potentially excluded relevant publications from other languages and from grey literature. Secondly, the analysis was conducted using the “pyBibX” library from Python and “bibliometrix” from R, both of which are powerful tools for bibliometric and thematic analysis. However, there are additional tools developed to extract valuable information from systemic reviews that could benefit our research using different approaches. Thirdly, while the chosen tools supported efficient data processing and visualization, they involved particular coding knowledge, which may restrict replicability for researchers without similar technical expertise. Future studies might benefit from broadening the range of tools and adopting more accessible approaches to improve the analysis’ comprehensiveness.

Lastly, preliminary stemming or lemmatization could be carefully applied to all the datasets to standardize word forms, reduce redundancy, and ensure consistency in text analysis. We only applied it so far for topic clusterization, Word Cloud, and N-Grams in Python in order to clean our data and eliminate similar groupings. Our main concern was not to overuse these technics since lemmatization can cause ambiguity for words having multiple meanings and can be time and resource consuming for very large datasets. Also, the use of the NLTK package depends on the WordNet Lemmatizer module that has a predefined set of English rules to be applied for this process. Stemming also can cause meaning loss due to overstemming, in which words may be reduced to the same root form by mistake (for instance, “policy” and “police” may both be reduced to “polic” even though they are unrelated in meaning), or understemming, in which, depending on the implemented algorithm and syntax rules, words that should be stemmed the same way are kept in different stemming groups. This is also affected by the language the algorithm is applied to, the strictness of stemming rules, the language slangs (ex. British, Australian, or American English), and the necessity to maintain context meaning or just keep distinct words. In the future, research can be improved by using more AI tools, such as integrating ChatGPT to answer questions based on the graphs obtained in the bibliographic analysis or by using more than two input sources, in addition to Scopus and WoS, customizing the use of indexing databases based on the publishing country or the author’s country, depending on what database is trending.

Also, a customized lemmatization (and stemming) function could be developed to remove critical redundancy for important family terms only, like digital or sustainable. This way, ambiguity is removed, and uniformization is applied at the same time.