I’ve been pretty silent the last few months here on my blog. I’ve been pretty busy with things like settling into our new house and starting full-time at Kynetx. A major sap on my free time both recently and for the last several years has been my Master’s Thesis. I’ve graduated now, and finally carved out some time to update my blog.

In my efforts in resuming regular blogging, I find it appropriate to post my Thesis for all the world. First, a little backstory.

For my thesis work, I worked on an algorithm to combine GPS tracklogs into what I call a spatial network. I chose this work because of my experience building ActiveTrails.com. As any excited graduate student (pre-thesis student, that is) I had grand ideas about what I was going to accomplish with my thesis work. Luckily for me, my graduate advisor guided me properly through the process, and I finally completed my work.

I do find it strange that only a written Thesis is required for an MS in Computer Science. I’ve decided that it only makes sense to post my code, that others might be able to experiment with my work without having to rewrite it from scratch. Now, I’m sure I’ve made plenty of mistakes in my code, and I hope that others can produce much better results then I, and not fall into the same lines of thinking that perhaps restricted my results.

I originally had plans to organize my code, clean it up, flush it full of comments, and organize my result files. And then I realized it might never happen. I’ve packaged my code, source files, and results into a zip file, and though it isn’t perfectly clean, I hope it’s useful for those who want to use it.

Progressive Spatial Networks: Learning from GPS Tracklogs (pdf link)

Source data, python source code, and result files (zip file).