Stack recommender is a website that uses machine learning algorithms to recommend Stack Overflow posts.. you can test drive it in recommender.im
- apis - use of stackoverflow's apis to extarct data
- util_sep - codes for handling the 20GB dump generated by stack overflow. Implementation of map reduce on single machine, mulltiple cores.
- Machine learning and data preparation scripts
- Flask web server that handles the requests to the machine learning API
- Data folder contains the trained model that needs will be used to predict posts. It is based on the user profile similarity
- Flask web server with the frontend that will find the user's id on stack overflow and call the apis, presenting the results to user.
- Ipython notebooks with data analysis. Check them rendered on http://nbviewer.ipython.org/github/adrianow/stackrecommend/tree/master/