Evaluating Financial Inclusion in Peru: A Cluster Analysis Using Self-Organizing Maps

Abstract

This study evaluates financial inclusion in Peru through self-organizing maps. Financial inclusion is a multidimensional issue of great importance on the global agenda and continues to concern various actors internationally. In this context, the objective is to assess the financial inclusion situation in the country and determine how self-organizing maps can complement standard models for this purpose. The empirical aim is to demonstrate how this technique can help identify priority areas and vulnerable groups, thus facilitating decision-making and policy design to improve the access to and use of financial services among Peruvian consumers by finding clearly defined profiles that allow the identification of potential problems within each category. This makes it possible to create customized strategies for each group, such as addressing the financial inclusion barriers faced by rural residents, compounded by low income and educational levels.

1. Introduction

Financial inclusion, defined as the ability of the population to access and utilize appropriate financial services, has become a key objective in socioeconomic policies over the past fifteen years, particularly within the agendas of developing countries (Sotomayor et al. 2018). In 2021, around 76% of the global adult population either possessed a financial institution account or utilized mobile money services, while nearly 71% of individuals in developing countries had banking access (Aurazo and Gasmi 2023). This growing interest is rooted in the belief that the involvement of all segments of the population in the financial system provides legitimacy and holds the potential for significant benefits to economic growth (Aurazo and Gasmi 2023). The significance of financial inclusion in developing nations is underscored by its crucial role in benefiting the impoverished and the most vulnerable members of society, fostering increased productivity and growth, and facilitating the transition to a digital economy. The resulting economic growth leads to increased income levels and a more equitable distribution of financial resources (Sotomayor et al. 2018).

While financial inclusion has historically been a challenge in Latin America and the Caribbean, there has been a gradual shift in recent years. Individuals in the region now have increased opportunities to access tailored financial products and services (Toledo Concha and León Reyes 2023). Nevertheless, despite the benefits offered by financial entities, a limited 39% of individuals seek loans due to insufficient financial market knowledge or access. Globally, about 50% of adults hold only one financial account, forcing many to resort to informal methods to fulfill their financial needs (Toledo Concha and León Reyes 2023). The low demand for financial services can be attributed to sociodemographic factors such as the lack of financial education and modest income levels, preventing individuals from fully leveraging economic access through traditional financial institutions (Toledo Concha and León Reyes 2023). Furthermore, recent research, such as that conducted in (Maehara et al. 2024), has focused on employing machine learning algorithms to assess financial inclusion in Peru. The conclusions drawn from this study underscore the significance of machine learning in predicting bank account accessibility and utilization among Peruvian consumers. The study identifies critical factors such as socioeconomic level, educational attainment, and age as significant indicators for bank account accessibility, with neural networks emerging as the most effective predictive model. In addition, the study highlights the importance of predictors such as using electronic devices, socioeconomic status, and income for predicting bank account usage.

In Peru, understanding the potential of financial inclusion encouraged taking various measures to improve access and use of supervised financial services, defining an adequate regulatory environment for the development of activities aimed at enhancing financial inclusion responsibly and sustainably (Sotomayor et al. 2018). Despite Peru being perceived as one of the most dynamic economies in Latin America, its recognition of financial inclusion as a policy priority, and its notable progress in that regard (Super Superintendencia de Banca Seguros y AFP (SBS) 2022), major disparities still exist in the levels of financial access and utilization across the country and it is still a developing economy characterized by a high degree of shadow economy, persistent labor informality, and increasing-but-low level of financial inclusion (Aurazo and Gasmi 2023). Significant variations are observed between Lima and other regions compared to neighboring countries and developed nations (Boitano and Abanto 2020). Moreover, a considerable portion of the Peruvian population still needs access to fundamental financial services, with 59% of adults aged 18 to 70 lacking a deposit account in the financial system, rising to 76% in rural areas (Sotomayor et al. 2018). Additionally, there remains a significant gap in financial inclusion, especially in rural and sparsely populated regions, where financial institutions often find it unfeasible to provide services with the existing channels and products (Super Superintendencia de Banca Seguros y AFP (SBS) 2022). In that sense, while financial inclusion has become a central concern for policymakers in Peru, significant efforts are still needed in data collection and theoretical development to address these issues effectively (Clamara et al. 2014). Accurate and high-quality information on financial inclusion trends is essential for developing, implementing, and assessing policies aimed at expanding the reach of the financial system (Super Superintendencia de Banca Seguros y AFP (SBS) 2022). As financial inclusion levels advance, the focus and data requirements will shift. Therefore, it is vital for regulators to periodically review the available data to identify information gaps and develop strategies to address them (Super Superintendencia de Banca Seguros y AFP (SBS) 2022).

2. Literature Review

2.1. Financial Inclusion

Financial inclusion has become a crucial element in global and national economic discussions, particularly concerning its role in the promotion of sustainable development. In essence, financial inclusion refers to the accessibility and availability of financial services to all individuals, especially the underserved. During the past decade, the push toward expanding access to financial services has led to significant advances in both traditional and digital inclusion efforts. In particular, traditional financial inclusion emphasizes access to banking services, while digital inclusion has grown rapidly with the advent of mobile money and digital payment platforms (Sha’ban et al. 2024). The integration of digital technologies into financial systems has enabled improved financial service distribution, lowered operational costs, and increased service efficiency, especially in emerging economies like China, where digital financial inclusion has proven instrumental in reaching underserved groups (Luo 2024).

In the given context, choosing to utilize or abstain from financial services is an integral aspect of financial decision-making. This process varies depending on generational differences, literacy levels, technological adaptability, and socioeconomic condition (Kumar et al. 2023; Saraswati and Kusumaningrum 2019). Research on financial inclusion has been widely distributed on both a global and local scale within individual countries. However, most of these studies typically offer generic insights into the country’s perspective and demographic groups based on variables usually analyzed broadly, without considering the specific characteristics in certain groups (Saraswati and Kusumaningrum 2019). Therefore, for a more comprehensive understanding of financial inclusion, it becomes crucial to scrutinize demand-side data: the emphasis should be on primarily examining the characteristics of both financially included and excluded groups when making financial decisions, based on identifiable groups with distinct behaviors and preferences taking into account various financial and socioeconomic factors (Saraswati and Kusumaningrum 2019). In this context, unsupervised techniques like clustering algorithms present themselves as attractive options for grouping individuals into similar categories based on their financial behaviors, sociodemographic characteristics, and other factors (Bester and Rosman 2024). One specific type of neural network particularly suited for cluster analysis is the self-organizing neural network developed by Kohonen (Jiménez et al. 2007). SOM networks offer an effective approach for exploratory data analysis, as they can categorize individuals with shared characteristics while also providing a visual representation of these groupings, which helps in pinpointing variables that are strong predictors for the phenomenon under study (Lee 2019). Such an approach would facilitate the formulation of specific recommendations, offering guidance to policymakers and strategies for financial regulators. This information would also empower the financial industry to undertake interventions that cater to the needs and raise awareness among the financially excluded groups, positioning them as potential users.

2.2. Machine Learning Clustering in Financial Inclusion

There is limited empirical evidence on the effectiveness of using machine learning methods to categorize individuals based on financial and socioeconomic factors with national data, evaluating variable importance and cluster heterogeneity in Peru. Furthermore, there have been no studies in Peru investigating the suitability of Kohonen’s self-organized maps for profiling individuals within the Peruvian financial system concerning financial inclusion and socioeconomic characteristics. Despite the increasing use of machine learning in many fields, there is little empirical evidence on the application of such techniques to categorize individuals based on financial and socioeconomic factors using national data in Peru. More specifically, no previous studies have explored the use of Kohonen’s SOMs to profile individuals in the Peruvian financial system with respect to financial inclusion and socioeconomic characteristics. This gap in the literature presents a unique opportunity to employ SOMs to provide new insights into financial inclusion in Peru, a country recognized for its efforts to promote inclusive financial systems (Clamara et al. 2014). From a microeconomic perspective, understanding the determinants of financial inclusion—such as income, education, and access to financial products—is critical for measuring the impact of these factors on participation in the formal financial sector (Aurazo and Gasmi 2023). However, traditional methodologies struggle to reveal how financial inclusion and sociobehavioral traits vary across a population or whether identifiable patterns exist within specific subgroups. This limitation makes SOMs particularly advantageous, as they allow for the detection of complex, non-linear relationships and the formation of distinct clusters within a population (Merzouki et al. 2021).

From a microeconomic perspective, there is a limited number of empirical studies that explore the determinants of financial inclusion and measure the impact of various factors influencing participation in the formal financial sector (Aurazo and Gasmi 2023). These studies highlight the importance of understanding the socioeconomic characteristics that influence how both households and businesses engage with the financial system. However, a significant drawback of these methodologies is their inability to demonstrate how financial inclusion and sociobehavioral traits differ across a population or if identifiable patterns exist within distinct groups (Merzouki et al. 2021). This study addresses these gaps by utilizing SOMs to identify specific financial inclusion profiles within the Peruvian population. Through the application of SOMs, the study uncovers subtle patterns and heterogeneities that traditional approaches may overlook. By highlighting these differences, the study provides a more granular understanding of the diverse factors affecting financial inclusion and offers actionable insights that can inform policy interventions and drive financial sector improvements. Using a clustering technique to compare and describe social groupings could test the hypothesis that sociobehavioral diversity accounts for the disparities observed within a country’s population.

Thus, this study aims to contribute to the literature by analyzing the application of Machine Learning techniques, such as Kohonen’s Self-Organizing Maps, to identify specific profiles within the population based on financial inclusion dynamics and sociobehavioral characteristics for the case of Peru, which is considered one of the best environments for financial inclusion among different developing countries around the world (Clamara et al. 2014). Considering the identified study gap, our paper addresses this by exploring the connection between financial inclusion, individual characteristics, and machine learning, proposing an ML approach that distinguishes itself through the use of SOMs; unlike previous studies that may have focused on broader or less detailed methods, this research emphasizes a granular approach, uncovering subtle patterns within the Peruvian population that have been largely unexplored.

This study assesses the utility of cluster analysis as an advantageous approach for delineating social clusters and aiding in state assessment measures in Peru, quantifying and characterizing agglomerations across the country. Specifically, the primary objective of this study is to develop and implement an approach based on Kohonen’s Self-Organizing Maps to uncover specific social patterns or profiles within the Peruvian population. This study distinguishes itself through its application of SOMs to the Peruvian context, an innovative use of this method that extends beyond conventional statistical approaches, allowing for the identification of complex patterns and non-linear relationships in heterogeneous datasets. This provides a more nuanced understanding of financial inclusion dynamics.

3. Materials and Methods

The National Survey of Demand for Financial Services and Financial Literacy (NSDFS), conducted in Peru between November and December 2022, served as the study data source (Super Superintendencia de Banca Seguros y AFP (SBS) 2022). The survey aimed to determine people’s knowledge, attitudes, and behaviors about financial matters. Furthermore, data on participants’ financial practices, economic resilience, and financial inclusion levels were gathered.

The sample was drawn from Peruvians 18 years and older living in urban and rural areas (23,375,001 individuals). A sample size of 1201 adults was determined to ensure national representativeness by sex and age, with a 95% confidence level and a margin of error of 2.8%. A multistage probabilistic sampling method was used, with strata defined by geographic region (Lima, Coastal, Highlands, Amazon) and area type (urban, rural). The surveys were conducted face-to-face from 25 November to 29 December 2022, in 24 departments and the constitutional province of Callao, covering 50 regions and 52 localities (Super Superintendencia de Banca Seguros y AFP (SBS) 2022).

The National Survey on Financial Services Demand and Financial Education carefully classified several sociodemographic characteristics to comprehend and examine the numerous elements that influence financial habits and inclusion among Peruvian residents. Each sociodemographic component taken into account in the study has a thorough operational definition given in

Table 1. Operational definition of the sociodemographic factors used in the study.

Variable	Definition	Scale	Categories
Accounts access (AA)	Individual has a savings account or a checking account or a time deposit account or savings in cooperatives	Nominal	1 = Yes 0 = No
Use of accounts in operations (UAO)	Conducted operations involving funds entering the account or funds exiting the account once a week or once a month over the past year	Nominal	1 = Yes 0 = No
Education Level (EL)	The highest level of education that the respondent has	Ordinal	1 = No education/Preschool 2 = Elementary school/Incomplete high school 3 = Complete high school/Incomplete technical education 4 = Complete technical education 5 = Incomplete higher education 6 = Higher education 7 = Postgraduate studies
Work or Occupation (W/O)	Occupation of the respondent	Nominal	1 = Employer 2 = Self-employed 3 = Technical/professional employee 4 = Laborer 5 = Domestic worker without salary 6 = Domestic worker with salary 7 = Member of the Armed and Police Forces 8 = Farmer/Countryman 9 = Dedicated to household 10 = Incapacitated for work due to illness or ill health 11 = Not working and not looking for it 12 = Student 0 = Other
Income (I)	Household’s monthly income of the respondent	Nominal	1 = <300 2 = 300–600 3 = 601–1200 4 = 1201–2400 5 = 2401–4800 6 = >4801
Gender (G)	Gender of the respondent	Nominal	1 = Male 2 = Female
Zone (Z)	Area of residence of the respondent	Nominal	1 = Rural 2 = Urban
Socioeconomic Level Response (SLR)	Socioeconomic status of the respondent	Ordinal	1 = SEL E 2 = SEL D 3 = SEL C 4 = SEL B 5 = SEL A
Socioeconomic Level Calculated (SLC)	Socioeconomic status of the respondent calculated by the interviewer	Ordinal	1 = SEL E 2 = SELC D 3 = SELC C2 4 = SELC C1 5 = SELC B2 6 = SELC B1 7 = SELC A2 8 = SELC A1
Use of Electronic Devices (UED)	The individual payments made for services once a week or once a month during the last year using smartphones or computers	Nominal	1 = Yes 0 = No
Age (A)	Age of the respondent	Ordinal	1 = 18–25 2 = 26–39 3 = 40–49 4 = 50+

. These variables include age, gender, place of residence, job/occupation, income, educational attainment, use of electronic devices for financial transactions, income, and socioeconomic position (as determined by the interviewer and reported by the respondents). The table lists the categories into which the respondents were classified and provides the description and scale for each variable. Using a structured approach, it is possible to gain a more nuanced understanding of how each element can affect behavior and financial literacy. This, in turn, enables targeted interventions to improve financial inclusion and education in various Peruvian demographic groups.

Self-Organizing Maps

The Self-Organizing Map (SOM) is a neural network that employs an iterative training method based on unsupervised, competitive, and self-organizing learning (Kohonen 1998).

Regarding their architecture, SOMs aim to map their weights to adapt to the input data by implementing a competitive layer. This layer is organized into a two-dimensional grid, where discrete nodal coordinates indicate the topological position of each node in the network (Haykin 2009; Kaski and Kohonen 1996; Kohonen 2013).

The algorithm of an SOM is primarily divided into two phases: competition and collaboration. During the competition phase, the Best Matching Unit (BMU), denoted as c, is calculated, which is the node closest to F, representing the input vector, in terms of the Euclidean distance to the weights w of each neuron:

$$\left|\right|F-w_c\left|\right|=\underset{i}{min}\left|\right|F-w_i\left|\right|$$

(1)

Subsequently, the weights of all nodes are updated using the learning rule:

$$w_i(s+1)=w_i\left(s\right)+\eta \left(s\right)h_{ci}\left(s\right)[F-w_i\left(s\right)]$$

(2)

where $\eta \left(s\right)$ represents the learning rate at iteration s, and $h_{ci}\left(s\right)$ is the neighborhood kernel around the winning node c. This neighborhood kernel decreases as the distance between node i and node c increases. To achieve convergence, both the learning rate and the neighborhood kernel radius decrease as the training procedure progresses.

It is crucial to note that while the competition phase identifies the winning unit, the true strength of SOM is manifested in the collaboration phase. During this phase, the influence of the winning node expands to its neighbors, allowing the network to adjust and organize coherently. Collaboration is achieved by gradually decreasing the learning rate ($\eta$) and the neighborhood kernel radius ($h_{ci}\left(s\right)$) as the training process advances. This phenomenon promotes convergence and the formation of meaningful topological representations, making SOM a powerful tool in the analysis and visualization of complex data (Haykin 2009; Kaski and Kohonen 1996).

In this study, SOMs were employed to identify distinct financial inclusion patterns within the Peruvian population. SOMs utilize competitive and self-organizing principles to map multidimensional data onto a lower-dimensional grid, making them particularly suitable for exploratory data analysis. One key advantage of SOMs is their ability to capture and preserve nonlinear relationships in complex datasets, allowing for the detection of underlying structures and trends that may not be immediately visible using traditional econometric techniques Vesanto and Alhoniemi (2000).

A Self-Organizing Map (SOM) not only stands out for its self-organizing ability but also for its application in data clustering. Once the training process reaches convergence, an additional phase is conducted to organize the data into coherent groups. In this step, the input vectors are evaluated in relation to the weights determined during training, using the central idea of assigning each vector to its BMU (Guthikonda 2005).

After convergence, each input vector is passed through the network again, but instead of updating the weights, each vector is assigned to its respective BMU. This proximity-based assignment process results in clustering, as similar vectors are grouped around neurons that best fit their characteristics. This approach creates natural groupings in which vectors that share similar attributes are clustered together in a core and are very similar to their immediate neighbors, while being distinct from those who do not share a neighborhood (Haykin 2009).

Furthermore, SOMs enable the preservation of topological relationships within the data, which means that similar data points are mapped closer together, enabling intuitive visualization of clusters and patterns. This feature is particularly beneficial in financial inclusion analysis, where the data might contain nonlinear relationships or hidden structures. By reducing dimensionality without compromising the integrity of the underlying data, SOMs offer interpretive insights into patterns of financial behavior that traditional models may overlook.

Essentially, the clustering phase leverages the SOM’s ability to organize its weights to reflect similarities in the input data. This process not only facilitates pattern identification and understanding of the intrinsic structure of the data but also provides a valuable tool for visualization and interpretation of complex datasets in diverse contexts.

The SOM’s ability to handle incomplete or noisy data also enhances its practical relevance in real-world applications, such as financial inclusion studies, where datasets often contain missing values. This robustness, combined with the clustering abilities of SOMs, allows for more dynamic insights into complex systems, further demonstrating their practical superiority in the context of evolving financial datasets. This study provides a novel perspective by using SOMs to uncover financial inclusion patterns in Peru, providing insights that would have been difficult to extract with traditional approaches.

The robustness of Self-Organizing Maps (SOMs) can be assessed using statistical tools designed to measure the confidence in their results. Specifically, quantization error and neighborhood relationships are employed to verify how well the data are organized within the map. Methods such as the bootstrap are utilized to provide reliable estimates of quantization error and the stability of the map’s topology, helping to determine whether the results are reproducible and significant. These approaches mitigate common issues like sensitivity to initial conditions and sample variability, making SOMs more robust compared to traditional neural networks De Bodt et al. (2002).

When compared to traditional techniques such as K-means clustering and Principal Component Analysis (PCA), SOMs offer several distinct advantages. While K-means effectively clusters data into a predefined number of groups, it lacks the ability to preserve topological relationships, which SOMs excel at. Unlike K-means, SOMs provide a two-dimensional representation that visually preserves the proximity of data points, making them more intuitive for interpreting clusters. On the other hand, PCA reduces dimensionality by transforming the data into principal components, but it relies on linear assumptions, which may oversimplify complex, non-linear data. SOMs, in contrast, can capture and organize non-linear patterns without such restrictions, making them more suitable for complex datasets like those encountered in financial inclusion studies. Thus, the use of SOMs in this study not only facilitates clustering and visualization but also ensures the preservation of essential data relationships that could be missed by more traditional methods (Bação et al. 2005).

4. Results

In the present research work, 25 distinct groups are identified applying the SOM that can be seen in Figure 1, which examines nine distinct attributes. The records are distributed along the topographic map, which allows one to determine the main characteristics of each group. The hexagonal grid used to represent the clusters in the figure shows how the number of components in each cluster is graphically represented by color intensity and numerical values. Darker shades indicate a higher concentration of elements, with numbers ranging from 3 to 115 in different clusters. Of the twenty-five groups, five were chosen: two with the highest presence of access to financial system tools, two with the lowest access, and the last group formed by women only with low access to credit.

As mentioned above, an analysis of the different groups was carried out. First, clusters 15 and 25 show a larger inclination to use accounts, as seen in Figure 2, according to the examination of the use of accounts of the various clusters. In particular, cluster 15 demonstrates a sizable proportion of users (62), while clusters 1 and 21 demonstrate very little to no utilization (cluster 1 has only 24 users and cluster 21 having 6 users). This implies that, in comparison to clusters 1 and 21, clusters 15 and 25 are more financially active. Furthermore, cluster 23, which is exclusively composed of females like cluster 25, has only 15 account users, suggesting little to no degree of account usage in this group unlike cluster 25.

Second, the account access information for several clusters is shown in Figure 3. Unlike account use, neither of the groups has a large portion of them using accounts regularly. The groups with more regular users are clusters 15 and 21.

Third, the information on credit access between different clusters is shown in Figure 4. We see that clusters such as 1 and 21 have many members without credit access, which could indicate structural issues or significant barriers for these groups. On the other hand, cluster 15 and 25, with an intermediate member count, show a better situation with regard to credit access. Cluster 23, despite being a smaller group, mirrors the credit access trends seen in cluster 1, indicating similar challenges despite the size difference between these groups.

Finally, gender distribution across the clusters is detailed in Figure 5. Cluster 1 consists entirely of men, totaling 115 individuals. In contrast, clusters 23 and 25 comprise solely women, with 75 and 89 members, respectively. Cluster 21 presents a more balanced gender mix, with 51 male members and 41 female members. This illustrates a pronounced gender bias in clusters 1, 23, and 25, while cluster 21 exhibits a more equitable gender distribution. Such disparities highlight the need to consider gender dynamics in financial inclusion strategies, ensuring equitable access to financial services across all demographics.

Profiles

Profile 1 (Cluster 1): This profile consists of 115 individuals, as can be seen in

Table 2. Profile 1.

Variable	Category	Accounts Access		Use of Credit		Accounts Use
		No	Yes	No	Yes	No	Yes
EL	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	26 (22.61%)	2 (1.8%)	26 (22.61%)	2 (1.74%)	28 (24.35%)	0 (0%)
	3	48 (41.74%)	16 (14.41%)	61 (53.04%)	3 (2.61%)	63 (54.78%)	1 (0.87%)
	4	11 (9.57%)	2 (1.8%)	13 (11.3%)	0 (0%)	12 (10.43%)	1 (0.87%)
	5	3 (2.61%)	3 (2.7%)	6 (5.22%)	0 (0%)	6 (5.22%)	0 (0%)
	6	3 (2.61%)	1 (0.9%)	4 (3.48%)	0 (0%)	4 (3.48%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
WO	0	3 (2.61%)	1 (0.87%)	4 (3.48%)	0 (0%)	4 (3.48%)	0 (0%)
	1	4 (3.48%)	2 (1.74%)	6 (5.22%)	0 (0%)	6 (5.22%)	0 (0%)
	2	60 (52.17%)	13 (11.3%)	70 (60.87%)	3 (2.61%)	71 (61.74%)	2 (1.74%)
	3	0 (0%)	2 (1.74%)	1 (0.87%)	1 (0.87%)	2 (1.74%)	0 (0%)
	4	8 (6.96%)	3 (2.61%)	10 (8.7%)	1 (0.87%)	11 (9.57%)	0 (0%)
	5	0 (0%)	1 (0.87%)	1 (0.87%)	0 (0%)	1 (0.87%)	0 (0%)
	6	2 (1.74%)	0 (0%)	2 (1.74%)	0 (0%)	2 (1.74%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	5 (4.35%)	1 (0.87%)	6 (5.22%)	0 (0%)	6 (5.22%)	0 (0%)
	9	1 (0.87%)	0 (0%)	1 (0.87%)	0 (0%)	1 (0.87%)	0 (0%)
	10	1 (0.87%)	0 (0%)	1 (0.87%)	0 (0%)	1 (0.87%)	0 (0%)
	11	3 (2.61%)	1 (0.87%)	4 (3.48%)	0 (0%)	4 (3.48%)	0 (0%)
	12	3 (0.87%)	0 (0%)	3 (0.87%)	0 (0%)	3 (0.87%)	0 (0%)
I	1	10 (8.7%)	1 (0.87%)	11 (9.57%)	0 (0%)	11 (9.57%)	0 (0%)
	2	26 (22.61%)	11 (9.57%)	35 (30.43%)	2 (1.74%)	36 (31.3%)	1 (0.87%)
	3	48 (41.74%)	10 (8.7%)	56 (48.7%)	2 (1.74%)	57 (49.57%)	1 (0.87%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	3 (2.61%)	1 (0.87%)	4 (3.48%)	0 (0%)	4 (3.48%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
G	1	91 (79.13%)	24 (20.87%)	110 (95.65%)	5 (4.35%)	113 (98.26%)	2 (1.74%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Z	1	70 (60.87%)	20 (17.39%)	87 (75.65%)	3 (2.61%)	88 (76.52%)	2 (1.74%)
	2	21 (18.26%)	4 (3.48%)	23 (20%)	2 (1.74%)	25 (21.74%)	0 (0%)
A	1	15 (13.04%)	5 (4.35%)	19 (16.52%)	1 (0.87%)	20 (17.39%)	0 (0%)
	2	22 (19.13%)	12 (10.43%)	33 (28.7%)	1 (0.87%)	33 (28.7%)	1 (0.87%)
	3	22 (19.13%)	3 (2.61%)	24 (20.87%)	1 (0.87%)	24 (20.87%)	1 (0.87%)
	4	32 (27.83%)	4 (3.48%)	34 (29.57%)	2 (1.74%)	36 (31.3%)	0 (0%)
SLC	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	7	91 (79.13%)	24 (20.87%)	110 (95.65%)	5 (4.35%)	113 (98.26%)	2 (1.74%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
SLR	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	91 (79.13%)	24 (20.87%)	110 (95.65%)	5 (4.35%)	113 (98.26%)	2 (1.74%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
UED	1	6 (5.22%)	7 (6.09%)	10 (8.7%)	3 (2.61%)	11 (9.57%)	2 (1.74%)
	0	85 (73.91%)	17 (14.78%)	100 (86.96%)	2 (1.74%)	102 (88.7%)	0 (0%)

. Within this group, 91 individuals (79.13%) do not have access to financial accounts; twenty-four people (20.87%) do have access but only two (1.74%) have used their accounts in the last month. Regarding credit access, the overwhelming majority, 110 individuals (95.65%), have not received loans from financial institutions, indicating minimal credit participation. The educational background of this group shows that a majority have up to complete secondary education, with 70 individuals (60.87%) reaching this level. The group consists predominantly of men, with 91 individuals (79.13%), and is primarily rural, with 80 individuals (69.57%) residing in rural areas. Occupationally, the group includes 34 self-employed individuals (29.57%), 12 farmers (10.43%), and others engaged in household duties. Most individuals in this profile earn a low income, with 80 individuals (69.57%) earning less than 1200 soles monthly. The technological involvement of this group is markedly low, with 81 individuals (70.43%) not using mobile phones or computers for financial transactions, reflecting significant barriers to digital financial inclusion.

Profile 2 (Cluster 21): This profile covers 92 individuals, as can be seen in

Table 3. Profile 2.

Variable	Category	Accounts Access		Use of Credit		Accounts Use
		No	Yes	No	Yes	No	Yes
EL	1	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)
	2	74 (80.43%)	5 (5.43%)	78 (84.78%)	1 (1.09%)	76 (82.61%)	3 (3.26%)
	3	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)	1 (1.09%)	1 (1.09%)
	4	7 (7.61%)	1 (1.09%)	8 (8.7%)	0 (0%)	8 (8.7%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	6	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
WO	0	6 (6.52%)	0 (0%)	6 (6.52%)	0 (0%)	6 (6.52%)	0 (0%)
	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	34 (36.96%)	3 (3.26%)	36 (39.13%)	1 (1.09%)	35 (38.04%)	2 (2.17%)
	3	1 (1.09%)	1 (1.09%)	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)
	4	4 (4.35%)	0 (0%)	4 (4.35%)	0 (0%)	4 (4.35%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	6	3 (3.26%)	1 (1.09%)	4 (4.35%)	0 (0%)	3 (3.26%)	1 (1.09%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	12 (13.04%)	1 (1.09%)	13 (14.13%)	0 (0%)	13 (14.13%)	0 (0%)
	9	18 (19.57%)	0 (0%)	18 (19.57%)	0 (0%)	18 (19.57%)	0 (0%)
	10	4 (4.35%)	0 (0%)	4 (4.35%)	0 (0%)	3 (3.26%)	1 (1.09%)
	11	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)	2 (2.17%)	0 (0%)
	12	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
I	1	42 (45.65%)	3 (3.26%)	45 (48.91%)	0 (0%)	42 (45.65%)	3 (3.26%)
	2	25 (27.17%)	1 (1.09%)	26 (28.26%)	0 (0%)	25 (27.17%)	1 (1.09%)
	3	9 (9.78%)	2 (2.17%)	10 (10.87%)	1 (1.09%)	11 (11.96%)	0 (0%)
	4	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)
	5	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)	1 (1.09%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
G	1	47 (51.09%)	4 (4.35%)	50 (54.35%)	1 (1.09%)	48 (52.17%)	3 (3.26%)
	2	39 (42.39%)	2 (2.17%)	41 (44.57%)	0 (0%)	40 (43.48%)	1 (1.09%)
Z	1	27 (29.35%)	3 (3.26%)	29 (31.52%)	1 (1.09%)	29 (31.52%)	1 (1.09%)
	2	59 (64.13%)	3 (3.26%)	62 (67.39%)	0 (0%)	59 (64.13%)	3 (3.26%)
A	1	6 (6.52%)	1 (1.09%)	7 (7.61%)	0 (0%)	7 (7.61%)	0 (0%)
	2	14 (15.22%)	1 (1.09%)	15 (16.3%)	0 (0%)	15 (16.3%)	0 (0%)
	3	19 (20.65%)	3 (3.26%)	21 (22.83%)	1 (1.09%)	22 (23.91%)	0 (0%)
	4	47 (51.09%)	1 (1.09%)	48 (52.17%)	0 (0%)	44 (47.83%)	4 (4.35%)
SLC	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	86 (93.48%)	6 (6.52%)	91 (98.91%)	1 (1.09%)	88 (95.65%)	4 (4.35%)
SLR	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	86 (93.48%)	6 (6.52%)	91 (98.91%)	1 (1.09%)	88 (95.65%)	4 (4.35%)
UED	1	7 (7.61%)	2 (2.17%)	9 (9.78%)	0 (0%)	5 (5.43%)	4 (4.35%)
	0	79 (85.87%)	4 (4.35%)	82 (89.13%)	1 (1.09%)	83 (90.22%)	0 (0%)

. Among these, seventy-nine individuals (85.87%) have access to financial accounts, although only six (6.52%) actively use these accounts for transactions. Credit access is similarly low, with only three individuals (3.26%) having taken loans from financial institutions, indicating a very limited engagement with credit facilities. Educationally, the group shows a varied level of attainment, with the majority having reached or surpassed secondary education. This profile includes 45 self-employed individuals (48.91%), suggesting a predominance of entrepreneurial activities within the group. It also highlights a significant number of people engaged in household duties or farming, reflecting a diverse occupational distribution. In terms of income, most people earn modestly, with a significant portion earning between 601 and 1200 soles per month. The technological use within this group is low, similar to Profile 1, with a majority not using mobile phones or computers for financial purposes, reflecting a gap in digital financial services adoption. Most of the individuals reside in rural areas, with 86 out of 92 (93.48%) based outside of urban centers. This demographic is also primarily male, with 51 males (55.43%) and 41 females (44.57%), showing a fairly balanced gender distribution.

Profile 3 (Cluster 15): This profile represents a group of 82 individuals, as seen in

Table 4. Profile 3.

Variable	Category	Accounts Access		Credit Access		Accounts Use
		No	Yes	No	Yes	No	Yes
EL	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)
	3	3 (3.7%)	4 (4.94%)	5 (6.17%)	2 (2.47%)	6 (7.41%)	1 (1.23%)
	4	7 (8.64%)	24 (29.63%)	22 (27.16%)	9 (11.11%)	30 (37.04%)	1 (1.23%)
	5	2 (2.47%)	13 (16.05%)	12 (14.81%)	3 (3.7%)	13 (16.05%)	2 (2.47%)
	6	6 (7.41%)	19 (23.46%)	21 (25.93%)	4 (4.94%)	25 (30.86%)	0 (0%)
	7	0 (0%)	2 (2.47%)	1 (1.23%)	1 (1.23%)	2 (2.47%)	0 (0%)
WO	0	0 (0%)	2 (2.47%)	1 (1.23%)	1 (1.23%)	2 (2.47%)	0 (0%)
	1	1 (1.23%)	5 (6.17%)	6 (7.41%)	0 (0%)	6 (7.41%)	0 (0%)
	2	4 (4.94%)	12 (14.81%)	12 (14.81%)	4 (4.94%)	15 (18.52%)	1 (1.23%)
	3	7 (8.64%)	28 (34.57%)	23 (28.4%)	12 (14.81%)	33 (40.74%)	2 (2.47%)
	4	0 (0%)	2 (2.47%)	1 (1.23%)	1 (1.23%)	2 (2.47%)	0 (0%)
	5	1 (1.23%)	1 (1.23%)	2 (2.47%)	0 (0%)	2 (2.47%)	0 (0%)
	6	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)
	7	0 (0%)	1 (1.23%)	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	9	3 (3.7%)	2 (2.47%)	5 (6.17%)	0 (0%)	5 (6.17%)	0 (0%)
	10	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)
	11	1 (1.23%)	9 (11.11%)	9 (11.11%)	1 (1.23%)	9 (11.11%)	1 (1.23%)
	12	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
I	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	1 (1.23%)	1 (1.23%)	0 (0%)	1 (1.23%)	0 (0%)
	3	2 (2.47%)	7 (8.64%)	6 (7.41%)	3 (3.7%)	9 (11.11%)	0 (0%)
	4	16 (19.75%)	29 (35.8%)	34 (41.98%)	11 (13.58%)	43 (53.09%)	2 (2.47%)
	5	0 (0%)	18 (22.22%)	15 (18.52%)	3 (3.7%)	18 (22.22%)	0 (0%)
	6	0 (0%)	5 (6.17%)	3 (3.7%)	2 (2.47%)	4 (4.94%)	1 (1.23%)
G	1	6 (7.41%)	18 (22.22%)	18 (22.22%)	6 (7.41%)	24 (29.63%)	0 (0%)
	2	13 (16.05%)	44 (54.32%)	44 (54.32%)	13 (16.05%)	53 (65.43%)	4 (4.94%)
Z	1	19 (23.46%)	62 (76.54%)	62 (76.54%)	19 (23.46%)	77 (95.06%)	4 (4.94%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
A	1	5 (6.17%)	22 (27.16%)	21 (25.93%)	6 (7.41%)	24 (29.63%)	3 (3.7%)
	2	6 (7.41%)	19 (23.46%)	18 (22.22%)	7 (8.64%)	25 (30.86%)	0 (0%)
	3	3 (3.7%)	12 (14.81%)	11 (13.58%)	4 (4.94%)	15 (18.52%)	0 (0%)
	4	5 (6.17%)	9 (11.11%)	12 (14.81%)	2 (2.47%)	13 (16.05%)	1 (1.23%)
SLC	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	6 (7.41%)	25 (30.86%)	21 (25.93%)	10 (12.35%)	29 (35.8%)	2 (2.47%)
	4	13 (16.05%)	37 (45.68%)	41 (50.62%)	9 (11.11%)	48 (59.26%)	2 (2.47%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
SLR	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	19 (23.46%)	62 (76.54%)	62 (76.54%)	19 (23.46%)	77 (95.06%)	4 (4.94%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
UED	1	10 (12.35%)	49 (60.49%)	45 (55.56%)	14 (17.28%)	55 (67.9%)	4 (4.94%)
	0	9 (11.11%)	13 (16.05%)	17 (20.99%)	5 (6.17%)	22 (27.16%)	0 (0%)

. Among them, 62 individuals (76.54%) have access to financial accounts and 19 (23.46%) do not. Despite the high rate of account access, only four individuals (4.94%) regularly use their accounts, indicating a low level of financial participation within the group. Credit access is similarly limited, with 19 individuals (23.46%) having received loans and 62 (76.54%) having no credit engagement.

Educationally, most people have higher levels of education, with 68 individuals (82.72%) completing secondary education or higher. Occupationally, a large part (40.74%) consists of technical or professional employees, while others are engaged in self-employment or household work. In terms of income, the majority (43 individuals, or 53.09%) earn between 1201 and 2400 soles per month, with smaller portions earning either less or more. The group is predominantly urban, with 95.06% of the individuals residing in urban areas. In terms of technology use, 55 individuals (67.90%) regularly use electronic devices for financial transactions, reflecting a higher level of technological integration compared to previous profiles. This profile represents a more financially engaged population, characterized by high access to accounts and urban residency. However, the low utilization of financial accounts and credit services suggests that, although access to financial services is present, barriers to active participation remain. This could indicate the need for more efforts in financial literacy or incentives to increase engagement with financial institutions.

Profile 4 (Cluster 25): This profile represents 89 individuals who are all women, as can be seen in

Table 5. Profile 4.

Variable	Category	Accounts Access		Credit Access		Accounts Use
		No	Yes	No	Yes	No	Yes
EL	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	3 (3.37%)	1 (1.12%)	4 (4.49%)	0 (0%)	4 (4.49%)	0 (0%)
	3	11 (12.36%)	25 (28.09%)	32 (35.96%)	4 (4.49%)	35 (39.33%)	1 (1.12%)
	4	7 (7.87%)	12 (13.48%)	16 (17.98%)	3 (3.37%)	19 (21.35%)	0 (0%)
	5	8 (8.99%)	7 (7.87%)	12 (13.48%)	3 (3.37%)	15 (16.85%)	0 (0%)
	6	1 (1.12%)	13 (14.61%)	13 (14.61%)	1 (1.12%)	14 (15.73%)	0 (0%)
	7	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)	1 (1.12%)	0 (0%)
WO	0	0 (0%)	1 (1.12%)	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)
	1	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)	1 (1.12%)	0 (0%)
	2	20 (22.47%)	31 (34.83%)	46 (51.69%)	5 (5.62%)	50 (56.18%)	1 (1.12%)
	3	6 (6.74%)	13 (14.61%)	16 (17.98%)	3 (3.37%)	19 (21.35%)	0 (0%)
	4	0 (0%)	3 (3.37%)	3 (3.37%)	0 (0%)	3 (3.37%)	0 (0%)
	5	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)	1 (1.12%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	9	1 (1.12%)	3 (3.37%)	4 (4.49%)	0 (0%)	4 (4.49%)	0 (0%)
	10	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	11	2 (2.25%)	3 (3.37%)	3 (3.37%)	2 (2.25%)	5 (5.62%)	0 (0%)
	12	1 (1.12%)	3 (3.37%)	4 (4.49%)	0 (0%)	4 (4.49%)	0 (0%)
I	1	1 (1.12%)	0 (0%)	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)
	2	1 (1.12%)	6 (6.74%)	6 (6.74%)	1 (1.12%)	7 (7.87%)	0 (0%)
	3	12 (13.48%)	11 (12.36%)	19 (21.35%)	4 (4.49%)	22 (24.72%)	1 (1.12%)
	4	14 (15.73%)	32 (35.96%)	44 (49.44%)	2 (2.25%)	46 (51.69%)	0 (0%)
	5	0 (0%)	5 (5.62%)	4 (4.49%)	1 (1.12%)	5 (5.62%)	0 (0%)
	6	0 (0%)	1 (1.12%)	1 (1.12%)	0 (0%)	1 (1.12%)	0 (0%)
G	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	30 (33.71%)	59 (66.29%)	78 (87.64%)	11 (12.36%)	88 (98.88%)	1 (1.12%)
Z	1	29 (32.58%)	54 (60.67%)	73 (82.02%)	10 (11.24%)	82 (92.13%)	1 (1.12%)
	2	1 (1.12%)	5 (5.62%)	5 (5.62%)	1 (1.12%)	6 (6.74%)	0 (0%)
A	1	4 (4.49%)	12 (13.48%)	13 (14.61%)	3 (3.37%)	16 (17.98%)	0 (0%)
	2	18 (20.22%)	37 (41.57%)	50 (56.18%)	5 (5.62%)	54 (60.67%)	1 (1.12%)
	3	4 (4.49%)	6 (6.74%)	8 (8.99%)	2 (2.25%)	10 (11.24%)	0 (0%)
	4	4 (4.49%)	4 (4.49%)	7 (7.87%)	1 (1.12%)	8 (8.99%)	0 (0%)
SLC	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	29 (32.58%)	58 (65.17%)	76 (85.39%)	11 (12.36%)	86 (96.63%)	1 (1.12%)
	6	1 (1.12%)	1 (1.12%)	2 (2.25%)	0 (0%)	2 (2.25%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
SLR	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	30 (33.71%)	59 (66.29%)	78 (87.64%)	11 (12.36%)	88 (98.88%)	1 (1.12%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
UED	1	11 (12.36%)	39 (43.82%)	40 (44.94%)	10 (11.24%)	49 (55.06%)	1 (1.12%)
	0	19 (21.35%)	20 (22.47%)	38 (42.7%)	1 (1.12%)	39 (43.82%)	0 (0%)

. The group exhibits moderate access to financial accounts, with fifty-nine individuals (66.29%) having access, although only one individual (1.12%) actively uses their account for financial transactions. This indicates a gap between access and utilization. Credit access is similarly limited, with only 11 women (12.36%) having accessed credit, while a significant majority, 78 individuals (87.64%), have not. Educationally, this group is relatively well educated, with a substantial number achieving higher education levels; 36 individuals (40.45%) have completed education beyond secondary school. The majority are employed in diverse fields, with a large segment self-employed or in professional roles, suggesting a mix of entrepreneurial and skilled occupations. Income levels show that many are in the middle income bracket, with 46 individuals (51.69%) earning between 1201 and 2400 soles per month.

Regarding technology use, 49 individuals (55.06%) use electronic devices for financial transactions, indicating a moderate level of technological adoption for financial purposes. Most of the profile’s individuals reside in urban areas, with 82 individuals (92.13%) living in such settings, which may facilitate their access to financial services. The entirely female group and urban residency suggest potential areas for targeted financial inclusion strategies, particularly to enhance active financial engagement and access to credit.

Profile 5 (Cluster 22): This profile includes 75 individuals who are predominantly female, as can be seen in

Table 6. Profile 5.

Variable	Category	Accounts Access		Credit Access		Accounts Use
		No	Yes	No	Yes	No	Yes
EL	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	15 (20%)	2 (2.67%)	15 (20%)	2 (2.67%)	17 (22.67%)	0 (0%)
	3	42 (56%)	12 (16%)	51 (68%)	3 (4%)	54 (72%)	0 (0%)
	4	2 (2.67%)	1 (1.33%)	3 (4%)	0 (0%)	3 (4%)	0 (0%)
	5	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
WO	0	3 (4%)	0 (0%)	3 (4%)	0 (0%)	3 (4%)	0 (0%)
	1	1 (1.33%)	2 (2.67%)	3 (4%)	0 (0%)	3 (4%)	0 (0%)
	2	23 (30.67%)	6 (8%)	26 (34.67%)	3 (4%)	29 (38.67%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	1 (1.33%)	2 (2.67%)	3 (4%)	0 (0%)	3 (4%)	0 (0%)
	5	2 (2.67%)	0 (0%)	1 (1.33%)	1 (1.33%)	2 (2.67%)	0 (0%)
	6	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	9	24 (32%)	3 (4%)	26 (34.67%)	1 (1.33%)	27 (36%)	0 (0%)
	10	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	11	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
	12	0 (0%)	1 (1.33%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
I	1	6 (8%)	0 (0%)	6 (8%)	0 (0%)	6 (8%)	0 (0%)
	2	13 (17.33%)	1 (1.33%)	14 (18.67%)	0 (0%)	14 (18.67%)	0 (0%)
	3	30 (40%)	12 (16%)	40 (53.33%)	2 (2.67%)	42 (56%)	0 (0%)
	4	4 (5.33%)	1 (1.33%)	4 (5.33%)	1 (1.33%)	5 (6.67%)	0 (0%)
	5	0 (0%)	1 (1.33%)	0 (0%)	1 (1.33%)	1 (1.33%)	0 (0%)
	6	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
G	1	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	2	60 (80%)	15 (20%)	70 (93.33%)	5 (6.67%)	75 (100%)	0 (0%)
Z	1	56 (74.67%)	13 (17.33%)	64 (85.33%)	5 (6.67%)	69 (92%)	0 (0%)
	2	4 (5.33%)	2 (2.67%)	6 (8%)	0 (0%)	6 (8%)	0 (0%)
A	1	9 (12%)	4 (5.33%)	12 (16%)	1 (1.33%)	13 (17.33%)	0 (0%)
	2	7 (9.33%)	1 (1.33%)	8 (10.67%)	0 (0%)	8 (10.67%)	0 (0%)
	3	18 (24%)	5 (6.67%)	19 (25.33%)	4 (5.33%)	23 (30.67%)	0 (0%)
	4	26 (34.67%)	5 (6.67%)	31 (41.33%)	0 (0%)	31 (41.33%)	0 (0%)
SLC	1	0 (0%)	1 (1.33%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	25 (33.33%)	5 (6.67%)	29 (38.67%)	1 (1.33%)	30 (40%)	0 (0%)
	6	35 (46.67%)	9 (12%)	40 (53.33%)	4 (5.33%)	44 (58.67%)	0 (0%)
	7	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	8	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
SLR	1	0 (0%)	1 (1.33%)	1 (1.33%)	0 (0%)	1 (1.33%)	0 (0%)
	2	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	3	60 (80%)	14 (18.67%)	69 (92%)	5 (6.67%)	74 (98.67%)	0 (0%)
	4	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
	5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
UED	1	3 (4%)	0 (0%)	3 (4%)	0 (0%)	3 (4%)	0 (0%)
	0	57 (76%)	15 (20%)	67 (89.33%)	5 (6.67%)	72 (96%)	0 (0%)

, with 60 individuals (80%) being women. This group demonstrates relatively limited financial access; 57 individuals (76%) do not have access to financial accounts and 67 individuals (89.33%) have not accessed credit facilities. Only three individuals (4%) have actively engaged in their accounts, pointing to a notable financial utilization gap. In terms of education, the majority of the group, 42 individuals (56%), have completed some form of secondary education but have not pursued higher education, which may influence their financial knowledge and access. The occupational distribution shows a predominance of individuals in lower-tier jobs, such as manual labor and informal sectors, with 23 individuals (30.67%) reporting that they are engaged in basic employment without technical or professional roles. The income distribution further underscores their economic position, with 40 individuals (53.33%) earning less than 1200 soles, placing them in the lower income bracket. This demographic exhibits limited use of technology for financial purposes, and a significant majority do not use mobile phones or computers for financial transactions (76%). Geographically, most people reside in rural areas, and 56 (74.67%) live outside of urban centers, which may further restrict their access to financial services. The age distribution shows a bias towards middle-aged adults, with 26 (34.67%) older than 50 years, potentially impacting their adaptability to new financial technologies and services.

5. Discussion

These five profiles represent groups with varying income levels, education, and professional activities. As shown in Figure 6, the most disadvantaged groups are those who lack access to banking services. These groups typically have lower levels of education and income, and a higher prevalence of individuals engaged in domestic tasks.

Profile 2 (Cluster 21) exposes the severe financial inclusion barriers faced by rural residents, compounded by low income and educational levels. This aligns with the findings showing that only 38.07% of rural households in Peru were included in the financial system, with the majority of rural residents earning less than PEN 251 per month and having incomplete secondary education (Quispe Mamani et al. 2024).

Profile 1 (Cluster 1) highlights restricted banking access among self-employed men, albeit with marginally higher incomes than those in Profile 2. The research supports this by noting that 72.18% of the heads of households are male, with greater financial inclusion in urban areas compared to rural areas, where self-employed men face similar challenges (Orazi et al. 2023; Quispe Mamani et al. 2024).

Profile 3 (Cluster 15) represents a better-situated group, with higher financial inclusion, education, and income, primarily composed of technologically connected males. Financial inclusion is strongly correlated with urban residency and higher education, as seen in 61.93% of urban residents having financial access, and higher educational levels significantly increase this probability (Quispe Mamani et al. 2024).

Profile 4 (Cluster 25) suggests financial stability with widespread access to banking, higher incomes, and participation in professional occupations. This is consistent with research showing that respondents who earn more than PEN 1256 monthly have a much greater likelihood of financial inclusion, indicating a strong link between income and financial stability (Orazi et al. 2023; Quispe Mamani et al. 2024).

Finally, Profile 5 (Cluster 23), mostly composed of self-employed women or those engaged in household tasks, faces significant educational and financial service access challenges. Women, particularly those in rural areas, were identified as facing notable barriers to financial inclusion, with only 27.82% of the respondents being women and indicating a clear gender gap in access to financial services (Quispe Mamani et al. 2024).

A comparison of the profiles reveals both similarities and differences in barriers and opportunities for financial inclusion. Profiles 2 and 5 share significant challenges related to low income and educational attainment, with both rural residents and self-employed women facing the steepest hurdles in accessing financial services. In contrast, Profiles 3 and 4 are similar in their relatively higher financial inclusion rates, benefiting from higher education levels, urban residency, and stable income. However, while Profile 3 consists primarily of technologically connected males, Profile 4 represents a broader group with professional occupations and widespread banking access, indicating that income stability is a key differentiating factor between the two. Profile 1, while experiencing restricted banking access, shows a more moderate situation with slightly better financial circumstances than Profile 2, reflecting the nuanced differences in financial inclusion between self-employed men depending on their income levels and urban or rural residence (Orazi et al. 2023; Quispe Mamani et al. 2024).

These profiles emphasize the diverse challenges across different demographic segments, highlighting the need for targeted interventions tailored to the specific needs of each group to improve financial inclusion.

In essence, future studies should involve closely analyzing biased or disproportionately high-weighting variables. The possible influence of these variables on group generation makes this precautionary measure highly advised, as they could unintentionally skew results and reflect real-life biases. These variables should be used cautiously, prioritizing those that could introduce bias in real-world scenarios. By carefully choosing variables, we can ensure that the group-generating processes are accurate and valid.

In Peru, the public and private sectors must work together to promote the availability of loans for those with demographic characteristics Profiles 1, 2, and 5. Through policy and regulatory frameworks that encourage financial inclusion initiatives, public institutions such as government agencies and development banks play a crucial role (Saha and Qin 2023; Song et al. 2024). These entities can enact policies that encourage financial institutions to extend their services to underserved rural areas (profile 1), ensure fair lending practices, and provide subsidies to support banks operating in high-risk segments. In addition, spending money on infrastructure to improve connectivity in isolated areas makes it easier to adopt digital banking solutions (Jejeniwa et al. 2024; Song et al. 2024), essential to reach those with little access to traditional banking services. Participation from the private sector is also necessary, especially from commercial banks, fintech companies, and microfinance organizations (Hammad Naeem et al. 2023). Commercial banks can offer customized financial products, such as microloans and low barrier savings accounts, to help low-income individuals and self-employed groups (Profiles 1 and 5). This is one way that commercial banks can innovate. Adopting mobile banking solutions broadens the reach of financial services to isolated regions (Senyo et al. 2023), improving efforts to close service delivery gaps in geography. Fintech companies contribute by creating digital platforms that allow peer-to-peer lending, mobile banking, and alternative credit scoring techniques (Cosma and Rimo 2023). These services are essential for individuals with traditional credit histories. Microfinance institutions enable people (Profiles 1, 2, and 5) to efficiently manage their finances and establish small enterprises by focusing on community-based lending and financial education (Song et al. 2024). Together, these measures can help public and private organizations improve financial inclusion in Peru, meet the specific requirements of various demographic groups, and promote sustainable economic growth throughout the nation.

6. Conclusions

In this study, we used a self-organizing map as a valuable tool to assess financial access in Peru. This method excels at dimensional reduction, allowing us to visually represent and understand complex data on the demand for financial services and sociodemographics in Peru. One of its main advantages is that it is an unsupervised learning technique that does not require labeled training data, making it ideal for exploratory data analysis. This is because self-organizing maps preserve internal relationships in the data by mapping similar data points to neighboring nodes, which is called topology preservation. In addition, self-organizing maps dynamically adapt to new data, making them well suited for evolving datasets and real-time applications.

However, certain limitations to this approach should be acknowledged. The interpretation of self-organizing maps can be challenging due to their unsupervised nature, which may require expert knowledge to analyze and draw meaningful conclusions from the visualized data accurately. Moreover, while self-organizing maps are powerful for exploratory analysis, they may not fully capture the underlying causal relationships between variables, which could limit the depth of insights derived from the data.

Future research could address these limitations by combining self-organizing maps with other analytical techniques, such as supervised learning models, to enhance the robustness and interpretability of the findings. Applying this method in different contexts, such as other countries or regions facing similar financial inclusion challenges, could further validate and expand its utility.

Despite these limitations, the insights gained from this study can potentially guide policymakers, stakeholders, financial professionals, and community advocates in creating focused interventions to advance fair access to financial services, encourage socioeconomic development, and include those who require it in the financial system. The study also provides a more comprehensive understanding of the variables that impact financial inclusion in Peru, which will aid future investigations and legislative initiatives. Furthermore, the study’s approach can be expanded upon and implemented in other nations or areas that encounter comparable obstacles concerning financial inclusion, thus enhancing its influence beyond Peru.