Research application: NY Times sentiment analyzer

Even though this project was a while ago, it marks my first encounter with computational social science, and therefore I still like it. The project showed me how fun computational methods are, and ultimately led me to reject my PhD grant and seek a position involving computational methods.

For the project, I built a simple, python/SQL-based application that lets users input a query term, and that outputs the NY times’ sentiments associated with this query.

These numerical values can be statistically compared to equivalent numbers from other newspaper sources, thereby tapping into biased reporting etc.

An example analysis using the application shows, for instance, that the NY times harbors more negative sentiments against Republicans than Democrats.

We had to do a video documenting our project. It was a pass/fail video that is unlikely to be watched, so it turned out slightly silly (timeless music etc.).