Building a Robust Web Scraper for Extracting Specific Address and Geographic Data from the Sucupira Platform: A Language-Agnostic Approach Using a Real Example

The Sucupira Platform, maintained by CAPES (Coordination for the Improvement of Higher Education Personnel) in Brazil, provides public access to detailed information about graduate programs (PPGs), including the section "Instituições de Ensino Superior Envolvidas" (Higher Education Institutions Involved). This section lists partner institutions with address components and geographic coordinates.

This technical article demonstrates the construction of a resilient web scraper focused on extracting the following fields for each institution: nome (institution name), logradouro (street/avenue), num (street number), complemento (complement – often absent), bairro (neighborhood), cep (postal code), ltd (latitude), and lng (longitude).

Using the concrete example page https://sucupira.capes.gov.br/sucupira/public/consultas/coleta/programa/viewPrograma.jsf?popup=true&cd_programa=41019016003P8 (which lists 4 institutions), we describe the observed HTML/textual structure and provide language-agnostic pseudocode. The approach prioritizes sequential line parsing over fragile selectors (IDs/classes), making it more resistant to minor layout changes.