In wjones127/thesis: Bike Routes Thesis (Title Case)

Given cyclists' commuting bicycle routes and their ratings of those routes, how can we determine which infrastructure features are not working for bicyclists? This is the question that first fascinated me when I started this thesis. As someone just getting into statistics, it was a mind boggling problem at the time, and even now I still don't have a complete answer; it turned out to be a larger research question than could fit in one undergraduate thesis. But here I make some substantial steps forward, refining the question ("how can we model how ride ratings are influenced by route, as well as weather, who is riding, and when they are riding?"), addressing part of the question (the non-route part), and coming up with some ideas on how to proceed further. \par This thesis represents a year's worth of hard work and the most enlightening educational experience I've had yet. It was a deep dive into how to do statistical modeling for a rich and complex data set with many open questions. During this undertaking, I couldn't be more thankful for the guidance of Andrew Bray, my thesis adviser. \par The story of how I got interested in this data was just as much a story of Reedies helping Reedies as it was of fascination with technical challenges. I first encountered Ride Report during Rennie Meyers' '15 presentation of her final project in the Introduction to Data Science class taught by Albert Kim in the spring of 2016. In a turn of luck, I found myself during the following summer interning at Switchboard (more formally known as Weathergram, Inc.), who shared an office with Knock Software, the creators of the Ride Report app. During that summer I got to talk with William Henderson '08 and Evan Heidtmann, the two developers at Knock, and became fascinated with the open questions in the data. I have to thank Rennie Meyers for sharing what she learned and Switchboard introducing me to William and Evan, but most of all thesis would not be possible without the generosity of William and Evan. They not only allowed me to examine the data they are obligated to keep as private as possible, but hosted me for hours in their office while I ran and debugged my models. \par Few feats at Reed are possible without the support of loved ones, and I've been especially fortunate here. I got through this semester in no small part because of the times I spend sharing great food with my friend Xian; the loving reassurance and support of my partner Kiki; and, of course, the love and financial support of my mom Kathleen.

wjones127/thesis documentation built on May 4, 2019, 7:34 a.m.