Sentiment Visualization for Twitter and Yelp

Inspiration

It's hard to find someone who doesn't participate in any social media (besides our mentor, Thanh), which is why we thought it would be thought-provoking to analyze _ how _ people are using social media. We did this by aggregating tweets and yelp reviews and analyzing them for sentiment and polarity.

What it does

Users can enter twitter keywords or hashtags, and our program generates a graph that shows sentiment towards those keywords/hashtags in terms of positivity and geographical location by displaying the results on a pie-chart and world map. As for the Yelp side, our program takes data from all of the current Yelp reviews and users and determines the polarity of the average review based on star rating. The results are then shown on three graphs (boxpot for polarity based on star ratings, histogram for polarity of a sample of overall reviews, line graph for change in polarity over the years based on star rating) for the user to use.

How we built it

The Twitter side of the project was created using Python and Javascript. The library TextBlob was used to do the natural language processing. The Yelp side of the project was built with Python using MatPlotLib and Seaborn.

Challenges we ran into

We are freshmen who have largely (almost completely) worked in Java during our programming careers, so transitioning into using Javascript and Python for this project was a big challenge. Another challenge we had was in linking all our separate work together without breaking any part.