|One hack to rule them all
|What the hack?!
Imagine the characters in Star Trek travelled in time to 2020. They discover this strange app called Twitter and are confused with everything that is going on. Now they want to help us, and the first part to solving a problem is identifying the cause of it. So they decide to build a model to visualise the cause of it all (inspired by LCARS in the original series)
What it does
The Twitter Content Access/Retrieval System (TCARS) allows the user to input a query like a keyword and then it extracts tweets from the past week that are relevant to it. Based on that data it creates multiple creative and useful visualizations like a world cloud, a co-occurrence graph and a polarity/subjectivity graph based on AI-powered sentiment analysis.
The polarity/subjectivity graph allows the user to visualize the general sentiments on a topic; relating how subjective the tweets are when compared against their dominant emotive opinion. The colour indicates the time of posting: ranging from 7 days ago represented by deep purple to the present day in bright yellow.
Polarity, on the x axis, lies between -1 and 1 representing a range of fully negative to fully positive sentiment. Subjectivity is represented on the y axis and measures the occurrence of words indicating opinion rather than fact – scoring 0 for stated facts and 1 for a fully subjective opinion.
As an example, a user could visualize the development of a controversial social issue with this software. Searching for keywords and observing how the sentiments of the tweets vary over time, one could quantitatively measure how the structure and tone of the conversation evolves.
How we built it
We used Python for both back-end and front-end implementation, more specifically we used TextBlob library for the sentiment analysis, regex for data cleaning and Pyglet for the user interface.
Challenges we ran into
We ran into a couple of technical challenges related to the user interface implementation and data cleaning but we successfully overcame them.
Accomplishments that we're proud of
We are proud of the way we faced the challenges, the knowledge that we gathered on the way and of our team dynamic.
What we learned
We learned more things about the intricacies of Twitter API, Pyglet and collaborative environments.
What's next for Twitter Content Access/Retrieval System (TCARS)
We plan to look for more interesting, creative and useful visualizations while maintaining the Star Trek theme.