SentiRed

This was the final project on the coding bootcamp at General Assembly. I pair programmed with a fellow student in one of the more academic and original projects I've worked on, namely a look at how machines can understand language. The Google Natural Language API offered sentiment analysis which we chose to run on content where we could be sure there would be a lot of sentiments, namely Reddit. This analysis was handled by the back end, which was written in Python using Flask and SQL to serve the data to the front end. The UI presented as a wrapper for Reddit, allowing you to look at a post and some of their comments. Each post you clicked on would be sent to the Google API along with the comments for sentiment anaylsis. This was then turned into an aggregate score and applied to the user's account.

Our next idea was to implement an emotional UI which updated according to the user's aggregate sentiment score. As a user's score went down, they would see harsher colours, pointier corners, less palatable fonts (Comic Sans being the worst). This dynamic theming persisted between sessions and gamified the Reddit experience. Hopefully for good.

NB The Google Natural Language API credits have expired since the project was created and an alternative is currently being worked upon. A full README is available in the GitHub repository.