Além do data warehouse: modernizando a governança analítica na Justiça Eleitoral

Abstract

The increasing demand for structured and accessible data within the Brazilian Electoral Justice system has driven the modernization of its analytical infrastructure. At the Regional Electoral Court of Rio Grande do Norte (TRE-RN), the existing architecture is based on a layered data model inspired by the medallion architecture and implemented using relational databases. Although functional, this model presents limitations in scalability, version control, and data traceability. To overcome these limitations, this paper proposes the implementation of a hybrid data architecture, known as a Data Lakehouse, using open-source technologies such as Apache Iceberg, Project Nessie, Apache Spark, Apache Airflow, and Dremio. Data is stored in an object storage system (Nutanix Objects) and made available through visualization tools such as Microsoft Power BI and Metabase. A proof of concept was developed in the TRE-RN environment, and the results demonstrated significant improvements in governance, versioning, and data traceability.