Uncovering Collaboration Patterns in Brazilian Computer Science Graduate Programs Through Network Embeddings

The Science of Science (SciSci) is crucial for understanding scientific production. However, there remains a gap in analyzing how geographic and productivity factors influence collaboration. This study explores collaboration networks in Brazilian Computer Science graduate programs. Using embeddings, clustering techniques, and Decision trees, we identified a larger group representing average scientific production, while smaller high-performing clusters often include researchers from prestigious institutions. Geographic disparities also highlight regional differences, with the South and Southeast regions dominating distinct groups. These findings emphasize the interplay between location, productivity, and impact, offering insights into collaboration dynamics.