This is a Python script that utilizes data mining, data visualization and the Twitter API. This script was developed for a Social Media and Data Mining course at Syracuse University, and I worked with a group of five other people. It was created in 2020.
To find the number/percentage of Twitter follower overlap for U.S. politicians in order to analyze how often people follow politicians of opposing political parties.
Although the intention is to compare politicians, any Twitter user can be used.
- Download the file to your computer.
- Open up the file in an IDE of your choice.
- Install any of the necessary libraries that are imported in this script using
pip install. - Get Twitter credentials here, and replace the fields (
CONSUMER_KEY,CONSUMER_SECRET,OAUTH_TOKEN,OAUTH_TOKEN_SECRET) with those credentials. - If you'd like, replace the current twitter users that are being used with your desired twitter users. You can identify the users with their screen name or their user ID. Modify the screen names at lines 152-161, and modify the user IDs at lines 140-150.
- Run the code. It will likely be stopped due to rate limiting and you'll have to wait 15 minutes for it to run again. This may happen more than once.
- The output will be a bar chart to visualize the comparisons (see common_twitter_followes.png), as well as a written summary with the percentages of the overlaps (see overlap_percentages.txt)
- With the exception of adding and removing a few comments, this code hasn't been updated since it was written in 2020. I have not optimized the code since then, but I may do so in the future.
- There are a lot of commented out sections within the code. I kept them there because they can be helpful should someone decide to use different parameters to get the data, or produce different visuals for the data.