Getting Sentimental: An Analysis of Twin Cities Tweets and Headlines

I worked with classmates to produce this sentiment analysis of local news headlines and tweets.

For our final project in a data science and visualization class, my group members and I analyzed newspaper content and tweets mentioning Twin Cities mayors. The result was a snapshot of local discourse in the Twin Cities over a roughly two week time span.

We created word clouds to give a general overview of the topics and sentiments defining local discourse. We then performed several types of text analysis and created visualizations to track tweet frequency, source, and other characteristics.

You can check out the full project here

This project helped me solidify several skills I had learned during the semester, such as scraping and wrangling data with RStudio.

Working on this project, one of the most salient lessons I learned was to think critically about the software that I’m using, and to acknowledge its potential for bias. For example, the sentiment analysis package we used identified the word “black” with negative sentiments, and the word “white” with positive sentiments. Especially when analyzing tweets and headlines related to Saint Paul’s first Black mayor, this categorization is highly inappropriate and problematic.