1. Pull relevant stock trading data from the internet
2. Create 2-4 models (use simple cookie cutter code mostly) to test performance of trading algorithm (regression, clustering, SVM, etc.) on back-testing data.
3. Generate a plot of portfolio performance/back test performance, and 4 other plots that are interesting/relevant (put in ipython notebook, with plots already outputted for me)
4. Entire code base should be less than 10,000 characters script (so not a very long script), in an ipython notebook I can run easily.
*Somewhat similar to code here (see these 6 parts explaining how to do a similar analysis, it includes all the code):
Or this is a simple version you can draw upon if needed:
***I need good comments for all major parts of the code as well, so I can understand it.
*Ideally you are modifying a project you have already completed, or something on the web, not doing this from scratch.