We were all keen to learn new web development technologies alongside Twitter's Streaming API.
What it does
- Takes a stream of random tweets using Twitter Streaming API
- Analyses the images in the tweets using Clarifai API
- Finds the most common things within sample photos
- Displays a live ranking of most common things
How we built it
We used Python and Flask, with Socket IO for a web socket. Too much JSON was involved.
Challenges we ran into
Clarifai doesn't work very well, the web socket was a nightmare. Tiredness can kill.
Accomplishments that we're proud of
We have a pretty decent end product, which works as intended and looks good.
What we learned
We learnt a lot about Python, Flask, web sockets, and the Twitter and Clarifai APIs.
What's next for TTTiPoT
Would work well with much more data: we were limited by the number of requests to Clarifai we could make. Implement more of Clarifai's models and potentially train our own model.