This course is taught by Andrew Ng. It provides the mathematical fundamentals for various machine learning techniques. The course provides a good amount of Matlab code for the student to use, but since I am already a Matlab expert I wanted to tackle the coursework using Python. I rewrote all the code from scratch using Python libraries in a Jupyter Notebook. Topics include linear and logistic regression, neural networks, unsupervised learning, and recommender systems.
The Fidelity Simple IRA allows the user to provide a selection of funds. The funds are listed in nested links and there are far too many for anyone to manually look through all of them. Fidelity also provides subjective ratings for various qualities of the funds like returns, costs, and safety. I wrote code to explore every link and pull the subjective ratings into a sortable database so we can make a decision on funds based on the most information possible.
One of my main hobbies is the trading card game Magic the Gathering. The story of the game is that players are wizards casting all the spells they know. Some spells summon lightning bolts to damage things and other spells summon grizzly bears to fight for you. These spells are associated with different color based on the land from which the wizard draws their power. The game has 5 different colors. This project is focussed on guessing the color of a creature card based on all of the other aspects of the card. I've applied many different techniques to this problem and all result in approximately the same accuracy. Currently I don't consider the text of the card, only the numberical data, but that is the next step and I think it would provide a large increase in accuracy.