Twitter Emotion/Sentiment Analysis

Our program when run on Elon Musk's Twitter

Inspiration

We wanted to do something hard, and in my honors they recommended a project similar to this, and I thought it would be fun. We thought it would be a good way to learn about applying our skills for the first time.

What it does

In theory, it is supposed to analyze a twitter users tweets, and understand the whether it is has positive or negative relationships with the related mentions and hashtags. This could be used to build connections between people and groups of people and better understand why people do what they do. It also can be used to know what relates people and what make them different. Sadly, due to time restrictions we don't have a UI and it has to be run through the command prompt.

How we built it

We used python. We have a parser that parses the tweets from Twitter, using Tweepy. Then we sent to an API called Indico to analyze the sentiment of the tweet.

Challenges we ran into

The parser was especially hard to implement. Tweepy would give us the data in non-string form, and we had trouble converting it to a string. Once we finally got past that we ran into the worst creation known to man-kind, EMOJI'S. These pesky rats wrecked havoc on the code everywhere. We couldn't find a good way to get past all of them so we just brute forced it and used try/catch to ignore them, which is probably a bad idea. Also, for some reason Tweepy will only get a few tweets, rather than all of them. For some reason Tweepy really hates hashtags and rarely gets the tweets with hashtags.