Turning a large bibliographic dataset into searchable, comparable research profiles.
A university team project integrating DBLP publication data with iCORE conference rankings and journal-ranking data. The ETL pipeline processes around 2.5 million publication records into a unified relational schema.
View repository ↗Data is cleaned and matched before the dashboard exposes search, profiles and comparisons through REST endpoints.
Cleaned and transformed publication records, modeled authors through many-to-many relationships, and matched venue names to ranking datasets. Publications without a venue match are retained so author and year analytics do not silently lose data.
Flask Blueprints organize conference, journal, author and year endpoints. MySQL connection pooling and SQL views support the dashboard’s analytical queries.
JavaScript, HTML/CSS and D3.js power autocomplete search, year-range filters, profile tables and interactive multi-venue comparison charts.
The pytest suite covers API contracts, ETL behavior, venue matching and frontend layout. GitHub Actions runs pytest and flake8 checks; API rate limiting and environment-driven debug configuration support safer application behavior.