We wanted a way to look at the United States in terms of happiness, and thought the platform that people express their emotions on most was twitter.
What it does
It analyzes tweets based on a library of words from twitter and their connotations to positivity or negativity.
How we built it
Using the Twitter API, we receive tweets and assign them either positive (10) or negative (1) values Each word from the tweets are then put into a pool, where it is then assigned a happiness value between 1 and 10 based on all of the tweets it was in (average sum of positive/negative) When a tweet is to be analyzed, its happiness score is taken from the average of its component words, and the final output is put on a scale of 1 - 10 (1 - 5 being more negative, and 5 - 10 being more positive)
Challenges we ran into
We ran out of time and had issues with streaming because of CORS, which meant that we weren't able to finish the end product, a map of the US with major cities displaying their average mood.
Accomplishments that we're proud of
Using training data on a large scale to create artificial intelligence on a computer is a great feat to accomplish in 12 hours.
What we learned
That CORS messes projects up and you need lots of sample data for AI to be accurate.
What's next for Twood
Creating the map of the US based on mood.