Plataforma de indicadores acadêmicos utilizando integração de dados por microsserviços

Abstract

Institutional scientific output is recorded in heterogeneous and dispersed sources, which makes the consolidation, analysis, and verifiability of indicators more difficult. This study presents the development and application of the Capivara Platform, a computational platform designed for the curation, integration, and interactive exploration of institutional scientific output. The proposed methodology organizes a workflow for the ingestion of institutional and curricular data, extraction and normalization of records from the Lattes Platform, traceable consolidation and deduplication of outputs, hierarchical bibliometric enrichment using external sources, and the provision of indicators with drill-down navigation from indicator to record and consistent export. Based on data extracted on January 24, 2026, the study analyzed 882 Lattes CVs of statutory faculty members at the Federal University of Triângulo Mineiro (UFTM), including retired faculty, totaling 169,657 production records in the 1971–2026 period. Persistent identifier enrichment increased DOI coverage from 7.17% to 8.58% in the overall scope and from 51.83% to 62.11% in the bibliographic core (articles, books, and book chapters), supported by external sources such as Crossref. Deduplication reached an approximate overall rate of 11.4% within the institutional affiliation scope, evidencing declarative redundancy and enabling comparative analyses across multiple facets (temporal, organizational, typological, collaborative, and bibliometric). It is concluded that the platform reduces the distance between aggregated indicators and their underlying records, providing a verifiable basis for exploratory diagnosis and support for the management of institutional scientific information.