Methodology

Background

I am working towards a model which could describes the causal relationships between neighborhood change, cycling infrastructure, and bike advocacy. The lack of relevant data makes analyzing those relationships challenging. Collecting additional data is necessary to be able to draw conclusions.

I began by using U.S. Census and American Community Survey (ACS) data to identify pairs of U.S. cities which have relatively similar demographics, but significantly different cycling rates. My initial study pairs Minneapolis, MN and Columbus, OH in the midwest, and Austin, TX and Charlotte, NC in the south. Travel funding provided the opportunity to visit these cities to speak with advocates and city officials, to gather data via bike and pedestrian counts, and to survey the road network.

Screenline counts

We are using the National Bike and Pedestrian Documentation Project (NBPD) methods for bicycle and pedestrian screenline counts. A diverse set of points in the road network are chosen based on known traffic, heatmaps from National Bike Month data, and input from local cyclists. Pedestrians and cyclists are counted for periods of two hours passing a specific point on the chosen road. NBPD suggests conducting counts at peak hours; however, as my premises include the idea that non-commuting utility cycling has increased, we will also conduct counts in restaurant districts in the evening hours.

Rolling surveys

To assess the nature and quality of the cycling infrastructure, we will use local cycling maps, Google Maps bicycle-identified routes, heatmaps from National Bike Month data, and input from local cyclists to determine routes to ride and survey. GPS tracks will be collected for all rides, and geotagged time-lapse video will be used to review the routes and identify issues.