Real Time Sentiment Tracking with Twitter

Code for the project can be found here.

Inspiration

Twitter is a social media platform where everyone posts their opinion on certain topics. Using sentiment analysis one can get a general idea of how the overall population is feeling towards a brand, company, political leader or anything else. By monitoring this information, companies can make more informed decisions that would improve customer satisfaction.

The Approach

The main two components of this web application are the streamer.py and app.py files respectively. In the streamer.py program, I make use of Tweepy, a simple Python library that uses the Twitter API to collect tweet data. Then, the sentiment of the tweets are extracted using VaderSentiment, a lexicon rule-based sentiment analysis tool, which is specifically tuned to perform well social media texts (i.e. Tweets from Twitter). All of this information is then stored into a local mySQL database.

The app.py program is in charge of data visualization. I use plotly and dash to create an interactive dashboard that presents all of the information. This program connects to the same mySQL database and processes the data in real-time.