PostGold

Post Gold

Everyone wants to get more likes and upvotes when making posts on social media, but it can be hard to think of new, creative content. To accomplish this we made a program that looks at data from posts that did well on Twitter and Reddit to help you create your own post that will achieve gold.

In making this we learned about the basics of natural language processing (NLP), Twitter and Reddit APIs, Markov Chains, and networkx (a python graph library).

This program is built with Django and deployed on a raspberry pi using the nginx web server and the gunicorn wrapper of uwsgi. We picked python as our primary language since it has good support for graphs and everyone on our team is comfortable with it.

The major setback that we faced in this project was the learning curve for natural language processing. It wasn't too hard to get basic sentences working and sounding like something that someone would actually post, but once we got to multiple sentences the algorithms begins to fail. We decided that the solution to this would be to implement a neural network but by the time we reached this decision it was too late to implement, so we settled for trying to improve your existing code as much as we could.

Overall the project was fun and we are proud of what we were able to produce in the time limit. If we continued the project we would look more into NLP and using a neural network.

Built With

• django
• gunicorn
• html
• nginx
• python
• uwsgi