The Media bias project was created in collaboration with the NAACP, WGBH and Boston University Spark. The NAACP and WGBH wanted to understand racial bias discrepancies in the media's written coverage. We took Boston as a case study and used local media sources - Boston Globe, WGBH, WBUR as data sources. The project produced analysis of thematic topics covering Black sub-neighborhoods using topic modeling, sentiment analysis, and volume of homicide coverage between Black and other racial groups.

I've managed multiple student teams on this project over its many iterations, the most recent iteration in development is building the tools into a ready -to-use web dashboard and improve the NLP components with the incorporation of more advanced models such as the BERT neural net.

View Code