A friend and I have been interested in combining our natural history and technology interests and one of the ways that has manifested is the idea of developing a bird mapping program. One of the grand ideas would be to be able to set out recording devices and have a program analyze the sounds to both identify and locate the position of the recorded birds. We haven't gotten too far with the idea yet, but periodically I dust the idea off and put some more work into it. This script is focused on locating the position of sounds (and is still a work in progress).

As an optional part of the UW data science course on Coursera, I started working on the Forest Cover Type competition on Kaggle. I ended up spending as much or more time on the competition than on the rest of the course because I was enjoying getting familiar with different machine learning techniques, exploring the data and trying to discover better ways of approaching the problem.