Indicadores de desempenho escolar: uma abordagem baseada em visualização de dados

Abstract

The pursuit of improvements in the quality of basic education in Brazil has been a constant challenge, particularly regarding the evaluation process, both in internal and external exams. The recent digitalization of schools has facilitated the collection of this information. After undergoing an Extraction, Transformation, and Load (ETL) process, these data enable the identification of relevant patterns, which are useful for decision-making. The use of Dashboards makes visualization more intuitive, facilitating interpretation by managers. By analyzing historical data from the institution itself, it is possible to identify evidence of below-expected performance in certain subjects, enabling the creation of new strategies to address these issues. In this context, this study aims to contribute to the improvement of public elementary education through Educational Data Mining and Learning Analytics, using interactive data visualization as a tool to support educational management in pedagogical meetings. The data source was the Integrated School Management System (SIGE) of a partner school, considering academic records of 9th-grade classes from 2013 to 2022. The analysis was conducted using the Python programming language and open-source libraries such as Pandas, Plotly, and Dash for data manipulation, transformation, and visualization. As a result of this work, interactive dashboards were developed for historical analysis, allowing the visualization of annual and bimonthly performance, the identification of seasonal patterns, and the extraction of relevant insights for the pedagogical coordination.