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

Week 3

This week was not a productive week for this thesis. But I will take this time right now to jot down some older thoughts I have had.

Within rider variation versus rider to rider variation

This is based on what I talked about with Andrew, my thesis advisor.

Obviously, different people will rate the same route differently. But there are different ways to model this. Here are three:

1. Assume the variation is simply due to varying street conditions. Thus, there is a probability that a person had a bad experience on a street, and this is independent from what person it is.
2. Assume the variation is mostly between riders. That is, riders will consistently give roughly the same rating for a given route, but different riders will have very different preference. This might mean there are different clusters of riders, such as those who don't like riding in traffic and those that are totally comfortable with that.
3. Assume the variation is mostly within riders. That is, there is no signifcance in the distribution one rider will rate a route than another rider.

The good news is that we can test on which of these assumptions is the most accurate. I will need to do some more testing to figure this out.

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