Recommends music to Twitter users based on the sentiment behind their tweet.
What it does
Analyzes the user's sentiment and mood through their text and recommends music based on that mood.
How I built it
The sentiment analyzer was built from scratch using scikit-learn, pandas, numpy and python. The training samples were downloaded from a large csv file containing tweets and pre-analyzed sentiments. This acts as a form of supervised learning for the model
The application is made with node.js with a mongoDB database hosted at mLab. The application can authenticate Twitter users using Passport.js.
Challenges I ran into
Learning machine learning from scratch. Trying to create an application by combining imcompatible languages and tools.
Accomplishments that I'm proud of
Being able to finish.
What I learned
Machine learning through scikit-learn, pandas, and numpy.
What's next for Moosik
Unique recommendations for users based on preferences and improved machine learning algorithm.