As we submitted a minute after (yes, after) the deadline we did not get a chance to elaborate on our simple yet sophisticated hack, so here is the more complete description:
Inspiration
With the recent scandals surrounding the spread of fake news, biased reports, and often outright erroneous information proliferating through social media, it's become important to be able to distinguish the objective from the subjective, something we humans are often unable to do due to our natural biases.
What it does
Collects recent posts from social media using Twitter's API (tweepy), runs the data through a language processing algorithm that determines the relative subjectivity and polarity of every post in multiple languages using TextBlob's Language Processing Library, then presents the data in an easy to read, simplified format using Plotly. The information is
Challenges we ran into
The biggest challenge was not being able to implement all the features we would have liked to see in the final product, as we really pushed our luck with the time limit, especially since we kept getting banned from our main APIs.
Accomplishments
We managed to build a functioning and useful application without any prior knowledge of Machine Learning and Data Representation libraries, as well as overcoming the hurdles of integrating multiple SDKs across different systems in the front and back-end
What we learned
All the above and team before machine :D
What's next for BS-Sense
Demographic filters, refine our text processing further, collect information from even more sources, present the data in a more sophisticated manner
Log in or sign up for Devpost to join the conversation.