The generation and management of household solid waste is a global problem that is aggravated by factors such as population growth, economic expansion (1). Li et al. (2) they anticipate an increase in waste generation in the coming years, considering that the global population already exceeds 8 billion inhabitants. Li and Wang (3) they highlight that this phenomenon enhances the production of domestic solid waste. Cheng et al. (4) and Pillay et al. (5) have laid a theoretical foundation by dealing with issues associated with the influence of psychological, social and economic factors on waste segregation behavior. In addition, significant theories have been recognized, such as the critical evaluation of studies related to moral and environmental education (6), and the relevance of humility as an essential virtue for environmental sensitivity (7).

The Ministry of the Environment MINAM (8) points out that in Peru 70% of municipal waste is produced by households; the newspaper El Peruano (9) It reports that during the year 2021, approximately 8 million 214355.90 tons of municipal solid waste were generated, which is equivalent to an average of 22505.08 tons per day. This waste was composed of 56.70% organic matter, 20.94% inorganic, 12.66% non-usable and 9.71% corresponding to hazardous waste. The newspaper Gestión (10) reports that MINAN has instructed the local authorities of San Vicente de Cañete, in Lima, to develop an action plan to address the problem of solid waste (SW) collection in that district.

Despite the efforts made, gaps persist in the understanding of the relationship between the PCG (Per Capita Generation) index and the production of domestic solid waste (DSW). This knowledge gap leads to poor solid waste (SW) management and inadequate infrastructure (3). In addition, the pandemic has altered the composition of waste and increased the demand for plastics (11). Requena-Sanchez et al. (12) point out that the lack of essential information to understand the dynamics of SW management hinders the implementation of strategies (13). They highlight the importance of SW management and care for the environment.

In this context, this study aims to fill these knowledge gaps and provide a more complete and detailed view of the problem of the generation of domestic SW in Peruvian households. To achieve this, we have resorted to the statistical approach for data evaluation with the support of the SEMMA data mining methodology, in order to improve the interpretation of data and thus optimize decision-making in SW management. In this sense, the main contribution of this work is to offer a data evaluation scheme that allows visualizing the production of SW and it is expected to explain the influence of the PCG index and the population level on the generation of domestic SW in Peruvian households, constituting a valuable tool for the visualization and targeting of policies to support the management of SW. offering a comprehensive approach to observe the distribution of domestic SW.

This technique consists of the stages: Sample, Explore, Modify, Model and Evaluate, established by the SAS Institute, to perform the essential tasks that facilitate the understanding, structuring, implementation and maintenance of data mining projects (14). Its implementation allowed us to structure a robust analytical process that addresses the heterogeneity and considerable volume of data associated with the generation of DSW. We also use principal component analysis (PCA) is one of the unsupervised learning strategies used in exploratory data analysis (15). In the case of DSW, variables such as SW composition, population levels and per capita generation rates present intricate relationships that make direct analysis difficult. The PCA not only simplified the solid residual data set by reducing it to principal components that capture the greatest explanatory variance, but also allows for the identification of latent structures and trends in DSW generation. The choice of SEMMA and PCA responds not only to their methodological robustness, but also to their alignment with the strategic objectives of the study, which seeks to close the knowledge gaps on the relationship between the PCG index and the production of DSW. Previous research has pointed out the difficulty of transforming complex data into practical tools for waste management (3, 12). Minan is committed to the 2030 Agenda for Sustainable Development (16) and our research is specifically aligned with the thirteenth Sustainable Development Goal which consists of taking urgent action to deal with climate change and its impacts.

The rest of this study is organized as follows: Section 2 presents the various related works that precede this study. Section 3 explains the adaptation of the SEMMA methodology of data mining used for a more effective interpretation of domestic SW production data. For the evaluation stage, only the interpretation of the results was considered, clusters were generated by the Ward method and PCA was performed for a better observation of the totality of solid waste. Section 4 contrasts the discussion between the results obtained and other similar studies. Section 5 sets out the conclusions.

Zhong et al. (17) proposes a set of general cybersecurity event analysis models combined with cybersecurity visualization and the classic SEMMA analytical model to standardize the analysis process. Network asset segmentation, anomalous network event extraction, and attack event evolution are implemented through multi-view synergy. The results show the validity of the analytical model. In 2017, China decided to introduce new regulations for the categorization of urban SW. Through a text analysis and a Measured Policy Coherence (MPC) model, it was concluded that none of the 11 policies evaluated were of high or low quality, obtaining an average of 7.71 in the MPC index (18). Popli et al. (19) construct a prediction model of urban SW in Laos from 1995 to 2050 using multilinear regression (MLR) with six socioeconomic parameters. Models are created in four scenarios and evaluated for accuracy with R 2 for performance. The model includes population and GDP per capita, it is the most effective for estimating the rate of waste generation. Municipal solid waste is projected to reach 0.98 million tons in 2030, 1.26 million tons in 2040 and 1.52 million tons in 2050. Likewise, Karadimas et al. (20) used the Genetic Algorithm (GA) to identify optimal routes in the collection of Municipal Solid Waste (MSW). Considering that between 60 and 80% of waste management expenditure goes toward collection, a small improvement at this stage can lead to considerable savings.

Fontaine et al. (21) developed predictive models for end-of-life waste management of products, underlining the relevance of detailed prior analysis; a case study is presented in Gatineau. This model allows the environmental impact of waste management strategies to be assessed, the data revealed a true standard deviation 84% higher than expected, calculating measures such as asymmetry 0.072, which indicates a symmetrical distribution. Similarly, Rifaldi et al. (22) conducted a study in Bekasi Regency, using a qualitative descriptive approach and causal loop diagram (CLD) techniques to understand the waste management system. The results indicate that the pandemic affects variables such as consumption patterns, economic conditions and the volume of waste, factors such as the number of people, public welfare and health, public consumption, waste volume, landfill area, waste management costs, technology, economic conditions and government budgets were identified.

In their research, Requena-Sanchez et al. (12) used an online form to gather and generate information on the production and characteristics of domestic solid waste in Latin America. It was found that 43% of respondents reported an increase in the total volume of their waste. The COVID-19 pandemic has exacerbated the use and disposal of personal protective equipment (PPE), an analysis in Lima, Peru, revealed the presence of 138 items of PPE on 11 beaches over a 12-week period. Recreational beaches were the most affected, followed by surfing and fishing areas. This phenomenon could be associated with the high density of visitors in these places. Most PPE was discarded by beach visitors (23). Likewise, the Community Leadership technique in Urban Environmental Sanitation was used in a small community in Malawi to evaluate its technical, political and economic feasibility. The findings showed that although water quality and access were satisfactory, the amount of water available was insufficient and waste management needed improvement (24).

Maalouf Mavropoulos (25) define solid waste as the sum of non-liquid and non-gaseous materials that are discarded, such as packaging, paper, plastic, glass, metal and food scraps. In the same vein, MINAM (8) refers to the waste generated in tons by cleaning actions in commercial, institutional and public spaces. According to the Pan-American Center for Sanitary Engineering and Environmental Sciences (CEPIS), 70% of municipal solid waste comes from homes, while the remaining 30% belongs to non-domestic waste of the total municipal waste. SW management involves all actions from generation to final disposal of SW. MINAM (8) examines the generation of waste per capita in Peru. Patrao and Karnik (26) highlight the importance of recycling awareness and the negative effects of poor e-waste management. Mumtaz et al. (27) in Poland demonstrate that proper management of Municipal Solid Waste can have a significant economic impact, and suggest wet oxidation as an effective technique to manage large amounts of SR. The reuse of waste in geotechnical engineering is also proposed as a solution to the growing production of solid waste due to urbanization and population growth (28).

Fernández-Domínguez et al. (29) 28 substrate/digestate pairs were analyzed in terms of organic RS quality. Using principal component analysis and hierarchical clustering, the results indicated a 58.02% variability in accessibility and a decrease in accessibility between 16 and 66% and an increase in complexity between 34 and 98%. In addition, an increase in non-biodegradable compounds between 17 and 66% was observed. This information was used for a model that predicts organic carbon mineralization in soil. Kim and de Vasconcelos Barros (47) analyzed the effects of the National Solid Waste Policy (PNRS) on populous municipalities in Minas Gerais, Brazil, using PCA. Two components were found to account for more than 70% of the variability of the data. The components were related to the objectives and principles of the PNRS. Zhang et al. (48) applied a novel electronic nose (EN) to extract information from the headspace gas of municipal solid waste incineration leachate. Combining multiple machine learning and PCA techniques, the EN-based UMAP-XGBT model had the best classification performance, with an accuracy rate of 99.95% in the training set and 95.83% in the test set. Demirarslan and Yener (30) in Turkey, where rapid population growth and urbanization have led to increased environmental challenges. The research analyzed municipal solid waste (MSW) generation, composition, disposal methods, and per capita waste production across different regions. Findings revealed an average MSW generation of 26.74 Mton/yr., with regional variations strongly correlated to population size. The Marmara region produced the highest waste volume, followed by Central Anatolia, Aegean, Mediterranean, Black Sea, South East Anatolia, and Eastern Anatolia. Population demographics explained between 84 and 95% of MSW variation. Also Demirarslan and Çelik (31) presents the first solid waste characterization analysis for Artvin city center in Turkey’s East Black Sea Region, examining waste composition across seven neighborhoods grouped by income levels (low, middle, and high). The research analyzed 11 waste categories, including organic, paper, plastic, glass, metal, ash, electronic, textile, garden, hazardous, and other materials. Using two-way ANOVA, researchers investigated seasonal variations in waste generation and its relationship with income levels. Results showed that organic waste dominated the annual composition at 61.06%, followed by paper (10.28%) and plastic (9%), with minimal electronic waste (0.037%). The study also evaluated three recyclable waste collection methods, concluding that building-based collection was the most economically viable option. The following questions were taken as questions for the study: What are the significant variations in the generation of household solid waste in the departments of Peru from 2014 to 2021? How does the population size influence the generation of household solid waste in the departments of Peru during this period? To what extent does the Per Capita Generation index influence the generation of household solid waste in the departments of Peru?

The methodological framework of this research was structured around the SEMMA methodology because of its ability to guide data mining processes in clearly defined stages. This methodology was selected for its flexibility and applicability in the analysis of large and heterogeneous data, such as the records of the SIGERSOL system. By adopting SEMMA, we sought to ensure a systematic approach to data preprocessing, transformation and analysis, thus allowing us to extract meaningful patterns related to the generation of DSW. The application of Principal Component Analysis (PCA) responded to the need to reduce the dimensionality of the data set while preserving as much variability as possible. This technique is especially appropriate in contexts where multiple correlated variables can make direct interpretation of the results difficult.

The analysis of the data collected and the use of statistical techniques corroborated that the number of population significantly influences the generation of DSW in the departments of Peru. Welch’s ANOVA and the square Eta provided solid evidence of a very large effect of the amount of population on the generation of DSW between departments. These findings are consistent with existing literature suggesting a direct relationship between population size and solid waste generation. As the population increases, so does the demand for goods and services, which in turn leads to an increase in waste production (38). Post hoc analysis using the Games-Howell test revealed statistically significant differences in the volumes of domestic SR produced among most of the pairs of population groups compared. However, there was a notable exception between the “1,000,000–1,189,000 inhabitants” and “1,222,000–1,538,000 inhabitants” groups, where no statistically significant differences were found in the amount of domestic SW produced. This may be due to factors not considered in this study, such as variability in waste management practices or differences in consumption levels. In harmony Quispe et al. (39) in the district of Santa Rosa, Ayacucho carried out an analysis showed significant differences between the SW. A high correlation between zones A and B indicates that organic waste exceeds inorganic waste.

On the other hand, Welch’s analysis of variance, Games-Howell’s post hoc analysis, and Student’s t-tests have provided solid statistical evidence that the PCG index per capita has a significant influence on the generation of domestic SW. The fact that the PCG accounts for 7.05% of the total variance in solid waste production levels between departments suggests that, although there are other factors that also influence waste generation, the PCG per capita has a moderate effect on household waste generation volumes. This statement coincides with what Lo Storto (40) found that the indicator was relative to the impact on eco-efficiency of variables that measure the size or density of DSW services, the indicators remain inconclusive.

Statistically significant differences in solid waste volumes were observed between the different groups of PCG. Groups with a higher PCG (“45.98–50.53%” and “46.0–50.99%”) generated more solid waste than groups with a lower PCG (“0–43.99%” and “44–45.99%”). Accordingly, Li et al. (2) the amount of waste is expected to increase in the coming years. The results of the study corroborate this theory, as they show a positive correlation between the different types of solid waste domestic SW, non-domestic SW and municipal SW and the number of inhabitants in the departments of Peru.

The PCA analysis and Ward’s minimum variance method revealed that Lima, being the department with the largest number of inhabitants, is also the one with the highest values in the production of domestic SW, constituting a group in itself. This observation suggests that solid waste generation is strongly influenced by population density. In agreement Cerna-Cueva et al. (41) due to the increase in population and consumption habits, the management of SW are environmental challenges in the world. In Group (a) the first component explained 81.5% of the inertia and the first two 99.4% of the variance. Lambayeque 5,616.48 tons, Loreto 2,946.44 and San Martín 1,596.07 had the highest DSW averages, while Amazonas 441.1 the lowest. Ucayali 60%, Loreto 58% and San Martín 57% registered the highest PCG indices. In Group (b) Ayacucho 701.81 tons showed the highest average DSW and Apurimac 497 the lowest. Tacna, Apurimac 44% and Moquegua 43% exhibited the highest PCG indices, while Huancavelica 42% and Pasco 41% the lowest. In Group (c) Piura 4,476.53 tons and La Libertad 3,478.46 with higher DSW averages, Huánuco 859.41 and Cajamarca 812.74 the lowest. Ica and Piura led with PCG of 48%, while Puno and Junín 43% were the lowest. A positive correlation was observed between DSW and number of inhabitants. These findings are consistent with those reported by Fernández-Domínguez et al. (29), who used PCA and hierarchical clustering analysis to characterize the quality of organic waste and found a 58.02% variability in accessibility, as well as significant changes in complexity and non-biodegradable compounds after treatment. Likewise, Kim and de Vasconcelos Barros (42) applied the PCA to evaluate the effects of the National Solid Waste Policy in municipalities of Minas Gerais, Brazil, and managed to represent more than 70% of the variability in the data. In addition, the study by Zhang et al. (43) demonstrates the usefulness of PCA in combination with machine learning techniques for the classification of leachate from municipal solid waste incineration, achieving accuracy rates of over 95% in training and test sets.

DSW generation patterns observed in Peru show notable similarities and differences with international studies (44). The high concentration of waste generation in Lima (13,220.47 tons) reflects a phenomenon documented in other Latin American megacities. Like Mexico City in 2020 (45), DSW is an issue of great relevance due to its environmental and social impact. According to recent data, the city generates approximately 13,149 tons of solid waste daily. This average translates to around 1.40 kg per inhabitant per day. This concentration pattern in major metropolises is consistent with trends observed in developing countries. Filipenco (46) menciona que la producción anual de DSW es más de 2.010 millones de toneladas de basura sólida urbana en todo el mundo. Si las tendencias actuales continúan, se espera que esta cifra aumente drásticamente y posiblemente alcance los 3.800 millones de toneladas en 2050.

Through the application of statistical techniques and the adaptation of the SEMMA data mining methodology and Principal Component Analysis (PCA) (47), it has been possible to obtain a more detailed and accurate view of the generation of domestic solid waste in Peru. This study has revealed the significant influences of both the amount of population and the PCG index per capita with the generation of domestic SW. the existence of a high positive correlation between domestic SW, non-domiciliary SW and municipal SW.

The t-tests give evidence of the effect of population size on solid waste generation (48). Groups with larger populations produced significantly more residues than groups with smaller populations, which is reflected in the large effect sizes found.

The PCG per capita has a moderate effect, which gives rise to further studies to further explore the underlying causes of these regional differences and evaluate the effectiveness of the waste management measures implemented to date (49).

The study indicates a positive correlation between the generation of residential solid waste (SW) and the number of inhabitants in the departments of Peru, which underlines the impact of the demographic volume on the production of domestic SW. In particular, Lima, with an average of 13220.47 tons, 59,813 residents and a PCG coefficient of 50%, is distinguished by its high production of residential SW, representing a unique case. PCA was performed for three groups as a result of grouping them by Ward’s minimum variance method, the results are detailed: In group (a) Lambayeque (5616.48 tons), Loreto (2946.44 tons) and Madre de Dios (2091.20 tons), in these the highest average in the production of domestic Solid Waste is shown. On the other hand, Amazonas (441.1 tons) and San Martin (1596.07 tons) exhibit the lowest averages. With regard to the per capita PCG index on average, Ucayali 60%, Loreto 58% and San Martín 57% have the highest values, compared to Madre de Dios 51%, Amazonas 52% and Lambayeque 53% have the lowest values. In group (b) Ayacucho (701.81 tons) presents the highest average of Domestic Solid Waste (DSW), while Apurimac (497 tons) registers the lowest number. In relation to the PCG indicator, the departments of Tacna with 44%, Apurimac with 44% and Moquegua with 43% exhibit the highest average percentages, in contrast to Huancavelica with 42% and Pasco with 41%, which display the lowest. In group (c), the departments with the highest average values of domestic solid waste are Piura (4476.53 tons) and La Libertad (3478.46 tons), while those with the lowest volume are Huánuco (859.41 tons) and Cajamarca (812.74 tons). With regard to the PCG, Ica and Piura lead with an average of 48%, as opposed to Puno and Junín, which have the lowest values with 43%.

The presence of Lima as a statistical outlier significantly influenced the analytical outcomes and subsequent interpretations. With an average DSW production of 13220.47 tons and a population of 59,813 inhabitants, Lima’s values deviated substantially from other departments, necessitating careful consideration in the analysis. This anomaly’s impact manifested in several key ways: Lima’s exceptional DSW production values necessitated the implementation of Ward’s hierarchical clustering method, resulting in three distinct groups and requiring separate PCA for each cluster. This methodological adaptation was crucial for maintaining the analytical integrity and preventing Lima’s outlier status from distorting the overall patterns in the dataset. Lima’s extreme values significantly influenced the dimensionality of the PCA, particularly in Dimension 1, where it emerged as the primary outlier. This affected the spatial distribution of other departments in the analysis, creating a distinct separation between high-production departments (Piura, La Libertad, Arequipa, and Lambayeque) and low-production regions (Madre de Dios, Huancavelica, Pasco, and Moquegua). Lima’s unique position in the dataset highlighted important demographic-waste generation relationships. Despite having a moderate PCG index of 50%, Lima’s total waste production far exceeded other departments. This finding particularly contrasts with departments like Ucayali, Loreto, and San Martín, which displayed higher PCG indices despite lower total waste production.

The study employs multiple statistical techniques Welch’s ANOVA, PCA, Ward’s method that complement each other, providing a comprehensive analytical framework. The use of SEMMA methodology adds structure to the data mining process. The high variance explanation 99.8% by the first two components of PCA demonstrates strong dimensional reduction effectiveness. The analysis provides robust statistical evidence of the influence of population on waste generation (Eta squared = 99.09%).

The extensive dataset (14,852 records from 196 registration forms) provides good coverage of all 24 departments of Peru. The study effectively captures regional variations in waste generation patterns. The study does not fully explore the underlying causes of the non-significant differences between population groups (1,000,000-1,189,000 and 1,222,000–1,538,000 inhabitants). The moderate effect of the PCG index (7.05% variance explanation) suggests unexplored factors affecting waste generation, providing valuable insights into Peru’s household solid waste generation patterns, particularly through its robust statistical approach. However, expanding the analysis to include more contextual factors and predictors would improve its practical utility for waste management planning and policies.

Domestic solid waste generation in Peru is highly correlated with population size, with Lima standing out as the largest daily generator of domestic solid waste, which requires specific management protocols. In regions such as Ucayali and Loreto, with high per capita generation rates of 60 and 58%, respectively, it is recommended to prioritize waste reduction programs. Analysis using Ward’s method identifies distinct regional patterns, highlighting the need for management policies differentiated by region. Furthermore, although the per capita generation rate moderately explains the variability 7.05%, it is crucial to investigate other factors that influence waste generation. Finally, in the face of projected waste growth, long-term infrastructure planning that integrates reduction and recycling strategies is urged to promote sustainable management.