Ciência de dados na análise de perfis e variáveis da insegurança alimentar em residências do Ceará

Abstract

The present work focuses on the analysis of food insecurity in families from Ceará, employing data science as the primary tool. The context addresses not only the identification of the problem but also the search for effective and targeted solutions through the application of advanced analytical techniques. Initially, an extensive dataset from the CMIC Survey (Characterization of Families in Extremely Vulnerable Social Situation), conducted by the Social Protection Secretariat of Ceará, was used. With approximately 34,000 families and 183 questions, the representative database was filtered to specifically focus on issues related to food security and demographics, totaling 29 essential questions. The adopted methodology encompasses the development and evaluation of an unsupervised machine learning model, with emphasis on the k-means algorithm. This clustering algorithm is applied to the socioeconomic, demographic, and behavioral data of families, identifying similar profiles regarding food insecurity. The model analysis includes performance metrics such as inertia, average distance from centroids, and silhouette, among others, providing a comprehensive assessment of the model’s ability to perform accurate groupings. The interpretation of the results seeks to identify key variables and patterns associated with the occurrence of food insecurity, providing valuable insights for the strategic allocation of public resources. The relevance of the study lies in contributing to the advancement of knowledge in the field of food security, presenting an innovative, machine-learning-based approach to identifying socioeconomic profiles in contexts of food scarcity. The obtained results not only have the potential to guide specific resources and actions but also offer assistance in decision-making to improve the quality of life for the most vulnerable families in Ceará. This work represents an effort to integrate modern technologies, such as machine learning, into the approach of complex social problems, exemplifying the role of data science in addressing real challenges.