Customer segmentation in online retail using k-means

K-means, Segmentação de Clientes em Varejo Online por Meio de

doi:10.22167/2675-6528-202602784

Article

September 30, 2026

Customer segmentation in online retail using k-means

Customer Segmentation in Online Retail Using K-means

Claudecir Réboli Junior; Henrique Raymundo Gioia

DOI: 10.22167/2675-6528-202602784

Article derived from a Final Course Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

In the context of the growing relevance of data for organizational decision-making, customer segmentation was analyzed as a strategic tool for value generation in online retail companies. The objective was to segment customers from a transactional database of an online store, aiming to identify consumption patterns and support marketing decisions. For this purpose, a quantitative approach was used, with the application of data treatment techniques, manual product categorization, and Z-Score normalization, followed by the implementation of the K-Means algorithm. The definition of the number of clusters considered the silhouette and elbow methods, complemented by an angular analysis. The results showed a highly concentrated structure, in which the majority of customers presented low individual value, while a small portion concentrated significant revenue share. The presence of unitary clusters was observed, interpreted as customers with extreme behavior and high strategic value. The analysis allowed for the identification of different consumption profiles and the proposal of distinct strategies, from scalable actions to highly personalized approaches. It was concluded that data-driven segmentation enabled a greater understanding of customer behavior and provided relevant subsidies for managerial decisions, highlighting growth opportunities and risks associated with revenue concentration.

Keywords: Data analysis; Consumer behavior; Clustering; Market segmentation; Retail.

1. Introduction

In the contemporary world, the consolidation of a data-driven decision-making culture has become a strategic objective for most organizations. However, raw data, by itself, does not support effective decisions; it is essential to subject it to treatment, organization, and analysis processes to convert it into relevant and consistent information (Herdiana et al., 2025).

The ability to transform data into valuable information is crucial for creating value in companies. Falcão et al. (2016) highlight that the value of information for an organization depends directly on the manager’s skill in analyzing and using it for correct decision-making, in order to achieve specific objectives.

In this scenario, customer segmentation emerges as a fundamental strategy. Kotler and Keller (2012) argue that dividing the market into segments is essential, as consumers have diverse preferences. John et al. (2023) reinforce that this strategy allows organizations to better understand and meet their consumers’ demands by grouping them based on shared characteristics. Haley (1968) adds that it is more effective to leverage existing market segments than to try to create new ones, adapting marketing strategies to consumer needs.

For market segmentation, several methodologies are employed, with the K-Means algorithm being one of the most frequent, as pointed out by Gomes and Meisen (2023) in a literature review. The popularity of K-Means lies in its simplicity and applicability to large databases. This method falls within the field of unsupervised machine learning, where algorithms are used to identify natural structures and groupings in data without prior supervision (Mahesh, 2019; Goldschmidt and Passos, 2005).

Despite the extensive use of customer segmentation techniques in the literature, a gap is observed in the practical application of these approaches in real transactional databases. This gap is particularly evident in the identification of consumption patterns that can be directly translated into effective business actions.

In this sense, the present research seeks to answer the following question: how can the application of the K-Means algorithm to online retail transactional data contribute to the identification of customer segments with strategic relevance?

The relevance of this study lies in its potential to provide an in-depth understanding of customer behavior, offering subsidies for managerial decisions and highlighting growth opportunities and risks associated with revenue concentration. Given this context, the present work aims to segment customers from an online retail transactional base using the K-Means algorithm, seeking to identify distinct consumption profiles and analyze their implications for the formulation of marketing, retention, and customer value maximization strategies.

2. Material and Methods

The present research adopted a quantitative methodological approach, with the purpose of segmenting customers from an online retail transactional base. The study focused on the application of data treatment techniques, exploratory analysis, and unsupervised machine learning algorithms, aiming to identify consumption patterns. This methodology was selected to answer the research question about the contribution of the K-Means algorithm in identifying customer segments with strategic relevance.

The empirical object consisted of a public transactional dataset, named “Online Retail”, obtained from the UC Irvine Machine Learning Repository. The database corresponds to transaction records of an online retail company, located in the United Kingdom, covering the period from December 2010 to December 2011. Initially, the set was composed of 541,909 records and eight variables: InvoiceNo, StockCode, Description, Quantity, InvoiceDate, UnitPrice, CustomerID, and Country.

The initial data cleaning and manipulation steps were performed in Microsoft Excel software, using the Power Query tool. Observations with null values or blank spaces in the CodCliente and Descricao variables were removed, as well as records where Quantidade was less than or equal to zero. Observations of CodEstoque that did not represent products (such as fees) were also excluded. After cleaning, the database was reduced to 396,370 observations.

Next, the manual categorization of products into categories such as Accessories, Food, Animals, Toys, Kitchen, Decoration, Lighting, Garden, Materials, Stationery, Romantic, Seasonal, Utilities, Glassware, Vintage, and Others was performed. This categorization aimed to structure the data for segmentation, ensuring consistency with the business context and reproducibility of the study (Falcão et al., 2016; Gomes and Meisen, 2023). After categorization, the transaction value per customer in each category was summed.

Customer segmentation was conducted through a computational approach based on Machine Learning, specifically unsupervised learning (Géron, 2019; Mahesh, 2019; Goldschmidt and Passos, 2005). The K-Means algorithm was implemented using the Python programming language, in the Spyder environment, through the Anaconda distribution. The choice of K-Means was justified by its simplicity and scalability on large datasets (Gomes and Meisen, 2023).

To circumvent the K-Means sensitivity to distinct scales, the data were normalized using the Z-Score technique (Goldschmidt and Passos, 2005). This normalization transformed the values of each attribute so that they presented a mean equal to zero and a standard deviation equal to one, according to the formula Z = (x – μ) / σ. The definition of the initial centroids and the parameters of the K-Means algorithm followed the guidelines of Johnson and Wichern (2007).

The settings adopted in the sklearn library for K-Means were: init = ‘k-means++’, max_iter = 100, n_init = 100, and random_state = 100. The determination of the optimal number of clusters, a critical step (Bassab et al., 1990), was performed using three methods: the silhouette coefficient (Géron, 2019), the elbow method (Géron, 2019), and a complementary angular analysis (Herdiana et al., 2025).

The silhouette coefficient calculated the average silhouette coefficient of all observations, using the formula (b-a) / max(a-b). The elbow method calculated the Within-Cluster Sum of Squares (WCSS), plotting the values to identify the inflection point of the curve, where WCSS = ΣΚ=1 ΣΧ₁€CK ||Xi – μκ||². Herdiana et al.’s (2025) angular analysis measured the angle formed between line segments of the elbow plot, indicating the smallest tan(ψ) value.

Although the silhouette coefficient and visual elbow analysis suggested two clusters as mathematically optimal, angular analysis indicated the formation of six clusters as the most suitable configuration. This decision was made to reconcile statistical optimization with managerial utility, allowing for the identification of more differentiated and relevant consumption profiles for the formulation of marketing, retention, and customer value maximization strategies.

3. Results and Discussion

The application of the K-Means algorithm to the transactional database resulted in the formation of six distinct clusters, according to the initial analysis. Cluster 0 stood out for grouping the vast majority of clients, with 4,298 individuals, representing 99.15% of the total base. This group was responsible for 6,093,130 of the transacted value, corresponding to 69.55% of the total revenue. In contrast, Clusters 1, 2, and 5 were composed of only one client each, representing a minimal share of the client quantity but contributing 3.19%, 2.96%, and 1.92% of the transacted value, respectively. Cluster 3, with 30 clients (0.69%), generated 1,388,548 of the total value, and Cluster 4, with four clients (0.09%), contributed 572,121 to the revenue. This initial distribution already indicated significant heterogeneity in the client base, with a concentration of value in a few individuals.

To validate the configuration of six clusters, an additional exploratory analysis was performed with the formation of four clusters. In this new configuration, Cluster 1 encompassed 4,299 customers (99.17% of the base), accounting for 6,261,602 of the transacted value, or 71.47% of the revenue. Cluster 0, with 31 customers (0.71%), contributed 1,505,278 of the value, or 17.18% of the revenue. Cluster 2, with one customer (0.03%), generated 259,657 of the value (2.96%), and Cluster 3, with four customers (0.09%), represented 734,529 of the total (8.38%). Although this four-cluster configuration might seem more simplified, it demonstrated a crucial limitation by grouping customer ID 16446, who had the fourth highest revenue in the base (168,472) and was highly specialized in the Stationery category, within a larger cluster. This merge would compromise the model’s ability to capture atypical and strategically relevant profiles for the business, justifying the maintenance of the original six clusters.

The decision to maintain the segmentation into six clusters, even with the presence of unitary groupings (Clusters 1, 2, and 5), aligns with the properties of the K-Means algorithm in databases with high heterogeneity and extreme observations. These unitary clusters were not considered statistical noise, but rather representations of customers with exceptional consumption patterns and high strategic value. Jointly, the three unitary clusters and Cluster 4 (with four customers) were responsible for 8.07% of the total revenue. The identification of these extreme profiles is managerially relevant, as it allows for the formulation of highly customized strategies, focused on customer retention and value maximization over time. The analysis of the centroids of each cluster by consumption category, after reorganization for greater clarity, revealed specific patterns that underpin marketing strategies.

Cluster Analysis

The cluster analysis was performed by grouping the results into three managerial groups, defined by similar characteristics regarding representativeness in the customer base, contribution to revenue, and consumption pattern. The first grouping, named Low-Medium Average Ticket, is composed exclusively of Cluster 1. This cluster concentrates 4,298 customers, corresponding to 99.15% of the total base, and is responsible for R$6,093,130 of the transacted value, or 69.55% of the revenue. The consumption profile of this group is diversified, with a higher participation in the Utilities, Kitchen, and Vintage categories, whose centroid values were 215, 146, and 134, respectively. The large number of customers and the dispersion of consumption across various categories indicate that this group is the main base supporting the online store’s sales volume.

Despite its vast representativeness in terms of number of clients, the participation of the Low-Medium Average Ticket group in revenue is proportionally smaller, suggesting a lower average transaction value per client compared to other groupings. From a managerial perspective, the main opportunity for this group lies in increasing the value generated per client. This can be achieved through scalable strategies, such as increasing the average ticket, encouraging repeat purchases, and implementing cross-selling actions. However, the massive size of the group limits the viability of highly individualized strategies, making it more effective to adopt large-scale segmented marketing campaigns, aiming to optimize engagement and conversion efficiently.

The second grouping, named High Average Ticket, is formed by Clusters 2 and 3. This group, although representing a small number of clients, with 30 clients in Cluster 2 (0.69% of the base) and four clients in Cluster 3 (0.09% of the base), has a high individual contribution to revenue. Cluster 2 accounts for 1,388,548 of the transacted value, or 15.85% of the revenue, with a higher concentration of consumption in Utilities (8,231), Materials (6,293), and Kitchen (5,726). Cluster 3, on the other hand, contributes 572,121 of the transacted value (6.53%) and shows a higher concentration in the Kitchen (19,699), Romantic (18,389), and Materials (17,828) categories. Together, these two clusters sum up to 34 clients, representing only 0.78% of the base, but are responsible for 22.38% of the total transacted value, highlighting their economic relevance.

The main characteristic of the High Average Ticket grouping is the significant economic contribution per client, making them strategic targets for retention and relationship maximization actions. From a managerial point of view, initiatives such as loyalty programs, personalized service, and up-selling offers are highly recommended for this group. The objective is to increase purchase recurrence and the long-term value of the relationship. However, the small number of clients in this grouping also represents a considerable risk for the business, as the loss of a few individuals can generate a disproportionate financial impact on the company’s total revenue, requiring continuous monitoring and care.

The third grouping, classified as Very High Average Ticket, is composed of Clusters 4, 5, and 6. This group is the smallest in the adopted solution, gathering only seven clients in total, which corresponds to 0.16% of the client base. However, these few clients are responsible for approximately 8.07% of the transacted value, demonstrating exceptionally high purchasing power and engagement. Cluster 4, with a single client, showed a strong concentration in Stationery, with a value of 168,470 in the centroid. Cluster 5, also with one client, stood out for Decoration (70,211), Romantic (31,822), and Vintage (31,132). Finally, Cluster 6, also unitary, presented a higher concentration in Utilities (53,998), Toys (30,485), and Materials (29,815).

The results for the “Altíssimo Ticket Médio” (Very High Average Ticket) cluster show consumption behaviors with extremely high transaction values compared to other groups. From a managerial perspective, these clients are considered of utmost economic relevance and require individualized and strategic monitoring. The opportunity lies in establishing long-term commercial relationships, with personalized service and integration with operational processes, such as logistics and supply planning, to ensure satisfaction and loyalty. However, the interpretation of these groupings requires caution, as the formation of clusters with few clients can be a result of genuinely differentiated consumption profiles or the influence of extreme observations on the K-Means algorithm. Thus, these findings should be viewed as evidence of extreme or highly differentiated behaviors in the studied database, and not necessarily as generalizable market segments.

Synthesis of Clusters

The organization of the results into three managerial groups revealed a consumption structure strongly concentrated in the analyzed customer base. The first group, encompassing the vast majority of customers, represents the main sales volume base and demands scalable marketing strategies. In contrast, the High and Very High Average Ticket groups, although numerically small, concentrate a significant portion of revenue, justifying highly personalized approaches focused on retention and individualized relationship management. This duality of consumption profiles underscores the need for a multifaceted marketing strategy capable of meeting the diverse needs and values of customers.

The concentration of revenue in a very small portion of the client portfolio represents a substantial risk for the business. Clusters 2 to 6, which correspond to only 0.85% of the client base, are responsible for 30.45% of the transacted value. This proportion is even more pronounced than the 80/20 relationship, often associated with the Pareto principle (Pareto, 1996), indicating a high dependence on a few clients for revenue sustainability. Therefore, the findings emphasize the urgency of developing differentiated relationship strategies, combining mass actions for the majority base with individualized, high-touch initiatives for high-value clients, mitigating risks and exploring growth opportunities.

Limitations of K-Means and Other Algorithms

Although the K-Means algorithm was selected for its simplicity, scalability, and widespread use in customer segmentation, its application in this study revealed some inherent limitations. K-Means relies on distances and centroids, making it sensitive to variable scale, data distribution, and notably, the presence of extreme observations. The identification of unitary clusters, while interpreted as high strategic value profiles, may also be a manifestation of this sensitivity to outliers, requiring careful interpretation of the results.

The literature points to the existence of alternative methods that could offer complementary perspectives. Hierarchical clustering, for example, would allow an analysis of group formation at different levels of granularity, without the need to pre-define the number of clusters, which could enrich the understanding of the data structure. Another algorithm, DBSCAN, would be useful for verifying whether the customers identified as outliers by K-Means are, in fact, isolated observations or part of denser regions. The absence of a comparison with these approaches constitutes a limitation of the research, as it prevents verification of the robustness and reproducibility of the identified clusters, leaving an opportunity for future investigations.

In summary, customer segmentation through the K-Means algorithm allowed for the identification of distinct consumption profiles and their strategic implications for online retail. The results revealed a heterogeneous customer base, with a large majority of low individual value and a small portion of very high-value customers, who concentrate a disproportionate share of revenue. This in-depth understanding of customer behavior provides valuable subsidies for formulating marketing, retention, and customer value maximization strategies, addressing both growth opportunities and risks associated with revenue concentration, and responding to the objective of identifying consumption patterns and supporting marketing decisions.

4. Conclusion

The study sought to segment customers from an online retail transactional database using the K-Means algorithm, aiming to identify distinct consumption profiles and analyze their implications for formulating marketing, retention, and customer value maximization strategies. A highly concentrated consumption structure was observed, where the vast majority of customers, corresponding to 99.15% of the base, presented low individual value but were responsible for 69.55% of the revenue. In contrast, a small portion of customers, representing 0.85% of the base, was identified as concentrating 30.45% of the total revenue, including single-unit clusters representing extreme and high strategic value consumption profiles. This segmentation enabled an in-depth understanding of customer behavior, providing subsidies for the formulation of differentiated marketing strategies, from scalable actions for the majority base to highly personalized approaches for high-value customers, and highlighted growth opportunities and risks associated with revenue concentration.

Despite the relevance of the findings, the research presented limitations. The K-Means algorithm demonstrated sensitivity to the scale of the variables, the distribution of the data, and the presence of extreme observations, which may have influenced the formation of unitary clusters. Additionally, the use of data from a single UK retailer and a one-year period restricts the generalization of the results, and the manual categorization of products may have introduced interpretative biases. For future studies, it is suggested to compare the identified groupings with alternative methods, such as hierarchical clustering or DBSCAN, in order to verify the robustness and reproducibility of the results, enriching the understanding of the data structure and managerial applicability.

Bibliographic References

Bussab, W. O.; Miazaki, E. S.; Andrade, D. F. 1990. Introdução à análise de agrupamentos. IME-USP, São Paulo, SP, Brasil. Disponível em: <https://cicerocq.wordpress.com/wp-content/uploads/2019/05/analise-de-agrupamento_bussab.pdf>. Acesso em: 12 abr. 2026.

Falcão, R. F.; Mazzero, S.; Campomar, M. C.; Toledo, G. 2016. A criação de valor para o cliente com base nos processos estratégicos de segmentação e posicionamento. Revista Interdisciplinar de Marketing, 6(2): 25–41.

Goldschmidt, R.; Passos, E. 2005. Data mining: conceitos, técnicas, algoritmos e orientações para aplicações. 1ed. Elsevier, Rio de Janeiro, RJ, Brasil.

Gomes, M. A.; Meisen, T. 2023. A review on customer segmentation methods for personalized customer targeting in e-commerce use cases. Information Systems and e-Business Management, 21: 527-570.

Géron, A. 2019. Hands-on machine learning with Scikit-Learn and TensorFlow: concepts, tools, and techniques to build intelligent systems. 2ed. O’Reilly Media, Inc., Sebastopol, CA, EUA.

Haley, R. I. 1968. Benefit segmentation: a decision-oriented research tool. Journal of Marketing, 32(3): 30-35.

Herdiana, I.; Kamal, M. A.; Estri, M. N. 2025. A more precise elbow method for optimum K-means clustering. arXiv (Cornell University).

John, J.; Shobayo, O.; Ogunleye, B. 2023. An exploration of clustering algorithms for customer segmentation in the UK retail market. Analytics, 2(4): 809–823.

Johnson, R. A.; Wichern, D. W. 2007. Applied multivariate statistical analysis. Pearson Prentice Hall, Upper Saddle River, NJ, EUA.

Kotler, P.; Keller, K. L. 2012. Administração de marketing. 14ed. Pearson, São Paulo, SP, Brasil.

Mahesh, B. 2019. Machine learning algorithms – a review. International Journal of Science and Research (IJSR), 9(1): 381-386.

Pareto, V. 1996. Manual de economia política. Editora Nova Cultural, São Paulo, SP, Brasil. Disponível em: <https://www.projetos.unijui.edu.br/economia/files/Pareto.pdf>. Acesso em: 12 abr. 2026.

Article originating from the Final Course Work of the Specialization in Data Science and Analytics of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform

Most recent

October 22, 2026

Desafios e Possibilidades do Primeiro Ano Escolar de Crianças com Tea para o Trabalho Docente

October 22, 2026

Jornada do Consumidor na Decisão de Compra de Infoprodutos no Ambiente Digital

October 22, 2026

Riscos Psicossociais e Saúde Mental no Trabalho em uma Instituição Bancária

October 22, 2026

Observatório Digital da Amazônia: Ciência de Dados Aplicada à Análise de Indicadores Sociais, Econômicos, Ambientais e de Saúde

October 22, 2026

Aprendizado Contínuo como Fator de Crescimento Organizacional na Percepção dos Colaboradores

October 22, 2026

Regressão Estatística em Microrredes para Estimativa de Desempenho Operacional: Caso Campusgrid

October 22, 2026

Previsão da Inadimplência Agregada no Sistema Financeiro Brasileiro por Modelo Sarimax com Variáveis Macroeconômicas

October 22, 2026

Fatores de Engajamento das Partes Interessadas na Transferência Internacional de Tecnologia para um Hospital Público

October 22, 2026

Combate às Perdas Comerciais de Energia Elétrica Utilizando Modelo de Séries Temporais

A inserção de crianças com Transtorno do Espectro Autista (TEA) na Educação Infantil apresenta desafios significativos, dada a dificuldade de adaptação em novos ambientes e a intensidade de estímulos. O objetivo foi analisar a atuação docente na Educação Infantil no processo de escolarização inicial de crianças com TEA. Realizou-se um estudo de caso, utilizando a entrevista narrativa como técnica de produção de dados. Os resultados revelaram que a atuação docente caracterizou-se por práticas de mediação e afeto, com a flexibilização da rotina e o uso de recursos lúdicos como estratégias essenciais para o desenvolvimento e organização sensorial da criança. Destacou-se também a importância da profissional de apoio escolar para o êxito do trabalho pedagógico. O desenvolvimento integral da criança foi promovido por meio de ações docentes intencionais que priorizaram a adaptação ambiental, o fortalecimento de vínculos e o atendimento às necessidades específicas do TEA. Como contribuição, elaborou-se um guia de estratégias pedagógicas voltado ao acolhimento de crianças com TEA em creches públicas.

Palavras-chave: Educação Infantil; Inclusão Escolar; Narrativas Docentes; Práticas Pedagógicas Inclusivas; Transtorno do Espectro Autista.

October 22, 2026

Jornada do Consumidor na Decisão de Compra de Infoprodutos no Ambiente Digital

A expansão do comércio eletrônico e a transformação tecnológica impulsionaram a necessidade de compreender o comportamento do consumidor de infoprodutos, cuja experiência percebida tornou-se crucial para o desenvolvimento profissional. O estudo objetivou analisar os fatores que influenciaram a decisão de compra de infoprodutos no ambiente digital, considerando as etapas da jornada do consumidor. Realizou-se uma pesquisa aplicada, quantitativa e descritiva, por meio de um levantamento de campo com 100 consumidores da Serra Gaúcha. A coleta de dados ocorreu via questionário estruturado, utilizando amostragem não probabilística. Os resultados revelaram um público qualificado e com forte comportamento “mobile-first”. Constatou-se que a disponibilidade de informações detalhadas e a aprovação social foram os principais gatilhos de atração na fase de pré-compra. Nas etapas transacionais e posteriores, a minimização de atritos operacionais, a agilidade do acesso imediato e a segurança sistêmica mostraram-se cruciais para a conversão e satisfação, embora lacunas na eficiência do suporte humano tenham sido identificadas. A retenção fundamentou-se na confiança gerada por experiências anteriores positivas, superando o fator preço. Concluiu-se que a jornada do consumidor de infoprodutos opera como um sistema integrado, exigindo alinhamento entre infraestrutura tecnológica, curadoria de conteúdo e humanização do suporte para assegurar o engajamento.

Palavras-chave: Comércio eletrônico; Comportamento do consumidor; Experiência do consumidor; Fidelização; Negócios virtuais.

October 22, 2026

Riscos Psicossociais e Saúde Mental no Trabalho em uma Instituição Bancária

A saúde mental no ambiente de trabalho é um tema de crescente relevância, impactando o desenvolvimento profissional e o bem-estar dos colaboradores. Um estudo foi realizado para identificar aspectos e situações que podem influenciar os níveis de absenteísmo em uma instituição bancária. A pesquisa envolveu 108 participantes e utilizou um questionário estruturado, baseado na versão curta do COPSOQ II de Gonçalves (2021), que continha 45 questões fechadas em escala Likert e duas perguntas abertas. Os dados foram coletados por meio de plataforma online e analisados com abordagem descritiva e quantitativa. Os resultados revelaram um cenário que indicou a necessidade de ajustes na distribuição das atividades e na percepção dos gestores sobre comportamentos e sinais de transtornos. Foram identificadas evidências importantes nas respostas abertas dos participantes. Concluiu-se que o estudo contribuiu para a promoção da percepção de um ambiente seguro para a saúde mental na organização.

Palavras-chave: Afastamentos por transtornos; Desenvolvimento profissional; Liderança operacional; Licença médica; Saúde mental.

October 22, 2026

Observatório Digital da Amazônia: Ciência de Dados Aplicada à Análise de Indicadores Sociais, Econômicos, Ambientais e de Saúde

O estado do Amazonas enfrentou desafios históricos na consolidação do Sistema Único de Saúde (SUS), decorrentes de barreiras geográficas, desigualdades socioeconômicas e limitações estruturais. Descreveu-se, neste estudo, o desenvolvimento e a implementação de um geoportal de informações em saúde, concebido para apoiar o monitoramento, a avaliação e o planejamento no contexto amazônico. Adotou-se uma abordagem de estudo de caso descritivo, baseada em análise documental e dados secundários, abrangendo o mapeamento de funcionalidades, a infraestrutura técnica e as etapas de validação da ferramenta. O Portal foi desenvolvido com tecnologias abertas e integrou múltiplas bases de dados nacionais, oferecendo dashboards interativos, mapas temáticos e repositório digital de documentos. Entre os resultados, destacaram-se a visualização georreferenciada de indicadores-chave, a possibilidade de análises comparativas temporais e a utilização da plataforma na revisão de planos de saúde e no redirecionamento de ações estratégicas. O portal pode representar um avanço na governança digital e na gestão baseada em evidências, com potencial de replicabilidade para outros territórios de alta complexidade, desde que acompanhada de investimentos em infraestrutura tecnológica, capacitação profissional e articulação interinstitucional.

Palavras-chave: Amazônia; Avaliação em Saúde; Saúde Digital; Sistemas de Informação em Saúde.

October 22, 2026

Aprendizado Contínuo como Fator de Crescimento Organizacional na Percepção dos Colaboradores

O impacto do aprendizado contínuo e da gestão do conhecimento no crescimento das organizações foi estudado em um cenário de globalização e constantes mudanças. O objetivo do estudo foi analisar a percepção dos colaboradores sobre a cultura de aprendizagem e sua importância para o desenvolvimento pessoal e profissional. Os dados foram obtidos por meio de uma pesquisa anônima, de abordagem quantitativa, utilizando um formulário com questões do tipo “survey” aplicado a 87 profissionais dos setores de tecnologia, saúde, educação, comércio e serviços. Os resultados indicaram que os colaboradores possuíam uma percepção positiva sobre a contribuição do aprendizado contínuo para o crescimento pessoal e organizacional, reconhecendo seus benefícios para o desenvolvimento de novas habilidades, inovação e competitividade. Contudo, observou-se que o incentivo à troca de conhecimento interno e a oferta de treinamentos práticos ainda apresentavam médias moderadas de concordância. Concluiu-se que o aprendizado contínuo contribui significativamente para o desenvolvimento individual e organizacional, sugerindo que as empresas devem desenvolver políticas estratégicas de aprendizado direcionado para novas habilidades, visando maior vantagem competitiva e inovação.

Palavras-chave: Aprendizagem contínua; Autodesenvolvimento; Crescimento organizacional; Gestão do conhecimento; Inovação.

October 22, 2026

Regressão Estatística em Microrredes para Estimativa de Desempenho Operacional: Caso Campusgrid

Um modelo de regressão estatística múltipla foi proposto para estimar o desempenho operacional da microrrede CampusGrid da UNICAMP, com base em grandezas elétricas internas. A microrrede integra geração fotovoltaica, baterias e um gerador a gás. Dados do sistema de gerenciamento de energia foram coletados em intervalos de cinco minutos durante um mês. O Índice de Confiabilidade Operacional, definido como a razão entre energia atendida e demandada, foi estabelecido como variável dependente. O modelo foi ajustado para o modo de operação ilhado, avaliando o impacto da demanda, estado de carga, geração solar e energia suprida. Os resultados revelaram elevado poder explicativo, indicando que a confiabilidade operacional depende da interação entre suprimento e demanda. Observou-se que o período noturno concentrou a maior taxa de atendimento parcial, atribuída à redução da geração solar. Concluiu-se que a regressão múltipla oferece uma estrutura interpretável e útil para o monitoramento de microrredes e o apoio a estratégias de controle.

Palavras-chave: Armazenamento em baterias; Geração fotovoltaica; Índice de Confiabilidade; Operação ilhada; Sistema de Gerenciamento de Energia.

October 22, 2026

Previsão da Inadimplência Agregada no Sistema Financeiro Brasileiro por Modelo Sarimax com Variáveis Macroeconômicas

A inadimplência agregada no sistema financeiro brasileiro é um indicador crucial de risco de crédito, cuja previsão é regulamentada pelas Resoluções CMN n.° 4.557/2017 e 4.966/2021 para a mensuração prospectiva de perdas esperadas. Modelou-se e previu-se a inadimplência agregada por meio do modelo SARIMAX, projetando sua trajetória até dezembro de 2026 sob cenários otimista, base e pessimista. Utilizaram-se 164 observações mensais (julho/2012 a fevereiro/2026) e variáveis macroeconômicas como desemprego, Selic, IPCA, IBC-Br e endividamento das famílias. A especificação com variáveis contemporâneas foi selecionada com base em critérios de informação e estabilidade dos coeficientes, e o diagnóstico dos resíduos confirmou adequação estatística, com MAPE de 3,12%. A inadimplência histórica variou entre 4,33% e 7,64%. A estimação identificou o desemprego (β = 0,2534; p < 0,001) e a Selic (β = 0,2128; p < 0,001) como principais determinantes positivos, com forte persistência temporal (AR(1) = 0,8965). A Selic exibiu correlação crescente com a inadimplência (r = 0,62 no lag 12), enquanto o desemprego demonstrou efeito contemporâneo (r = 0,33). As projeções indicaram inadimplência entre 6,30% e 7,98% em dezembro de 2026, com uma amplitude de 1,68 ponto percentual, evidenciando implicações diretas para o dimensionamento de provisões nas instituições financeiras. Palavras-chave: Desemprego; Modelagem preditiva; Política monetária; Provisão de perdas; Séries temporais.

October 22, 2026

Fatores de Engajamento das Partes Interessadas na Transferência Internacional de Tecnologia para um Hospital Público

Este estudo descreveu os fatores críticos percebidos para o engajamento das partes interessadas na transferência internacional de tecnologia em um hospital público, com foco no planejamento. O objetivo foi descrever esses fatores condicionantes, conforme a percepção de profissionais. Realizou-se um estudo de caso único, aplicando um questionário a 25 profissionais e utilizando análise de conteúdo, complementada por tabulações descritivas. Os principais obstáculos identificados foram resistência à mudança (72%), financiamento (64%), falta de comunicação (60%) e burocracia (40%). Apenas 28% dos respondentes reconheceram a existência de ferramentas adequadas para gestão de stakeholders, enquanto 48% negaram e 24% não souberam informar. Cruzamentos exploratórios indicaram variações na avaliação de influência conforme o tempo de atuação, mas não estabeleceram relações causais. A falta de engajamento foi associada a atrasos, baixa adesão, desperdício de recursos, falhas de comunicação, comprometimento da absorção de conhecimento e riscos à segurança do paciente. As conclusões restringiram-se às percepções do grupo pesquisado e fundamentaram recomendações preliminares para comunicação e coordenação intersetorial.

Palavras-chave: Estudo de caso; Gestão de projetos; Inovação em saúde; Partes interessadas; Setor público.

October 22, 2026

Combate às Perdas Comerciais de Energia Elétrica Utilizando Modelo de Séries Temporais

As perdas comerciais de energia elétrica representaram um dos principais desafios para as distribuidoras, impactando diretamente a sustentabilidade econômica e a qualidade do fornecimento. Avaliou-se a eficiência de modelos de séries temporais, com foco em técnicas de suavização exponencial, na previsão do consumo faturado de unidades consumidoras, para apoiar a identificação de desvios associados a perdas não técnicas. Utilizaram-se dados históricos de consumo mensal de uma distribuidora de energia elétrica no Brasil, segmentados por perfis de comportamento (estacionariedade, tendência e sazonalidade). Aplicaram-se modelos de suavização exponencial simples, Holt e Holt-Winters (aditivo e multiplicativo), comparados a uma média móvel tradicional. O erro percentual absoluto médio (MAPE) foi o critério de avaliação. Adicionalmente, empregou-se a abordagem ETS para seleção automática do modelo e testes de diagnóstico dos resíduos, como Ljung-Box e análise da função de autocorrelação. Os resultados demonstraram que modelos mais robustos apresentaram maior capacidade de adaptação às características das séries e menor erro, com destaque para o desempenho superior em cenários com tendência. Para séries sazonais, o modelo ETS mostrou maior capacidade de adaptação a transições. Concluiu-se que a utilização de modelos de séries temporais mais estruturados contribuiu significativamente para a melhoria da acurácia das previsões e ofereceu suporte analítico relevante para a identificação de perdas comerciais e otimização de processos operacionais e comerciais no setor elétrico.

Palavras-chave: Distribuição de energia; MAPE; Otimização; Previsão de consumo; Suavização exponencial.

Economy

October 19, 2026

End of the 6×1 scale: what do the numbers allow us to conclude?

Following the promulgation of the end of the 6×1 shift system, the debate on the shorter workday moves to implementation and requires distinguishing social benefits from economic costs. The end of the 6×1 shift system transitioned from legislative discussion to a constitutional change in Brazil. On October 16, 2026, the National Congress promulgated Constitutional Amendment No. 141/2026, originating from Proposal […]

Artificial Intelligence

October 16, 2026

Why study if AI exists?

When answers become easy, learning comes to mean more than just arriving at them. Why do I need to learn this if I can ask an artificial intelligence? This type of question has accompanied education for decades. Why learn an operation if a calculator exists? Why memorize information if it can be found […]

Tax Management

October 15, 2026

Pink tax and the gender of price

How different marketing treatments for men and women are incorporated by consumption taxation Imagine a shelf on which two razors of the same brand and intended for the same purpose are displayed side by side: one has dark packaging and promises strength, precision, and efficiency; the other comes in shades of pink or […]

Health

October 13, 2026

The missing link in health transformation

Why is innovation not enough to drive changes in organizations and health systems Digital transformation has become a strategic priority for health systems worldwide. Electronic health records, artificial intelligence, telemedicine, remote monitoring, and data analysis are now part of the strategic agenda of governments and health organizations. Paradoxically, […]

Marketing

October 09, 2026

Bitcoin or pets: what Artificial Intelligence would not see

A real-life consumer intelligence case shows what gets left out of the brief A few years ago, an American insurance company hired a specialized consumer intelligence firm to answer three questions: The insurer was also considering launching paid insurance in bitcoin. At the time, the consultancy’s workflow was as follows: the analyst created a query […]

School Management

October 06, 2026

The school is not a casino: when trading cards become currency in the playground

Why, for certain students, does school manage to offer so few experiences of achievement? The World Cup is over. Brazil left the competition, the school holidays came, the final was played and the football calendar moved on. It would be reasonable to imagine that, with the end of the tournament and the return to classes, […] would also begin

Energy Transition

September 30, 2026

Renewable energies: the Brazilian advantage

Brazil needs to invest in the capacity to transform natural resources into innovation, productivity, and development. Talking about energy is talking about development. Behind industry, transportation, services, cities, and increasingly artificial intelligence, there is an energy infrastructure that sustains the economy. Understanding a country’s energy matrix is understanding […]

Data Analysis

September 30, 2026

When the data changes, has reality changed?

An analysis of almost 20 million teaching positions-year between 2020 and 2025 shows how changes in population composition and record production can alter the interpretation of historical series Between 2020 and 2025, the number of formal teaching employment positions identified in RAIS (Annual Relation of Social Information), from the Ministry of Labor, […]

Artificial Intelligence

September 28, 2026

Artificial intelligence and the trends

Why do so many people join in on the fun with ChatGPT? I was never the trend person. I still am not. Some pass through my timeline, I see them, I find them interesting but not to the point of trying them out. This was the case, inclusive, with some recent waves of images made by artificial intelligence, like the 80s trend, the one that […]

Tax Management

September 01, 2026

Tax reform and estate planning in the countryside

Efficiency in managing taxes facilitates governance and allows for the organization of family succession. Rural wealth organization has become a strategic measure in light of the tax reform in Brazil, which came into effect this year. The planning of activities and rural wealth management, through tools such as the transition from an Individual to […]

Renata de Gaspari Valdejão Almeida

Artificial Intelligence

August 05, 2026

Inside the company and off the radar: the risks of Shadow AI

According to Microsoft, 75% of professionals already use artificial intelligence at work without company approval. The use of generative artificial intelligence is already part of the corporate routine. Tools like ChatGPT, Copilot, and other platforms help professionals produce texts, analyze data, create presentations, and accelerate tasks in practically all sectors. But this advance […]

Renata de Gaspari Valdejão Almeida

Strategy

July 02, 2026

The future as a competence

Academic interest reinforces the projection of scenarios as an essential tool for decision-making. While some companies are still trying to get the next step right, others are already working with multiple possible futures — and preparing for all of them. But after all, what does it mean to “think about the future”? Is it a trend, a strategy, or a new competence that […]

Renata de Gaspari Valdejão Almeida

Innovation

Technology

May 21, 2026

Agro SP Corridor: innovation, institutional articulation and future

Pecege commands the executive operation of the 400 km ecosystem activities that became known as the Brazilian agro Silicon Valley In 2025, in addition to taking over the management of the Piracicaba Technology Park, the Pecege Institute was also designated to command the executive operation of the State Agribusiness Innovation Corridor activities, the so-called […]

Education

April 27, 2026

How to turn a TCC into a business

New management of the Piracicaba Technology Park offers consulting to MBA graduates The new project director of the Piracicaba Technology Park (PTP), Alexandre Barreto, often says that one of the greatest rewards he has ever had in life was seeing his MBA USP/Esalq students structure companies from their final projects […]

Technology

March 20, 2026

The human repertoire in the age of AI

Critical thinking and creativity are irreplaceable in the work of a marketing professional Artificial Intelligence (AI) is capable of incredible feats, but it does not yet possess the judgment, empathy, intuition, and strategic creativity necessary to do the job of a good marketing professional. In this area, it is important to know how to use the tool to automate […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

February 23, 2026

ESG as a growth strategy in a cooperative in RS

Santa Clara balances social and environmental responsibility with innovation Founded in 1911 by Italian immigrants in Carlos Barbosa, in the Serra Gaúcha region, Cooperativa Santa Clara is an example that it is possible to balance economic growth with social and environmental responsibility. From a small cheese production company, the organization has evolved into a […]

Renata de Gaspari Valdejão Almeida

Agribusiness

Innovation

January 22, 2026

Brazil free of Foot-and-Mouth Disease: what changes for Brazilian beef?

With a herd of over 200 million head of cattle, the country is currently the leader in exports. Brazil recently received, for the first time, the certification of country free of foot -and-mouth disease without vaccination, conferred by the World Organisation for Animal Health (WOAH). The seal strengthens the country’s position as a reliable supplier of meat for the […]

Renata de Gaspari Valdejão Almeida

Innovation

December 09, 2025

Innovation: the new cycle of the Piracicaba Technology Park

Instituto Pecege assumes management, betting on the creation of new businesses Starting in 2026, Instituto Pecege will assume the management of the Piracicaba Technology Park (PTP). Among its responsibilities will be the administration, operation, maintenance, and expansion of the park, in addition to fostering activities in innovation, entrepreneurship, and science and technology. The choice was made […]

Renata de Gaspari Valdejão Almeida

July 06, 2026

Large-Scale Architecture: DSR-SCRUM Change Management

Mônica Alessandra Guerios; Maria do Carmo Assis Todorov Abstract generated by ResumeAI, an artificial intelligence solution developed by Instituto Pecege focused on synthesis and writing. Architecture and urbanism are configured as essential bases in an increasingly complex and challenging urban space, where the intrinsic connection between social, environmental, and technological issues […]