MODELING BICYCLE RIDERSHIP: CHALLENGES, OPPORTUNITIES, AND SCALABLE SOLUTIONS FOR NORTHWEST ARKANSAS

This study addresses the growing trend of bicycling in urban settings and the challenges in accurately estimating bicyclist counts. Its broader ambition relates to improving the data and information base to increase and diversify the cycling ridership of communities across the United States. By improving upon and expanding the knowledge base on accurately counting ridership within a large geographic area over time, we aim to enhance the methods and analyses of past, present, and future studies on this topic. The easier and more accessible these counting methods can become for communities to adopt, the greater the chance of success in promoting ridership in these areas for widespread use to represent the diverse characteristics of each location and provide more socially equitable options for commuting. As a researcher who has personally experienced the benefits of cycling in NWA, it is evident how it can improve the quality of life for people across all demographics and ease the environmental burden by using less motor transport and building a stronger sense of community through this shared enjoyment of cycling. Our methodology involves a thorough comparative analysis of data from various sources, then adding covariate and location-specific data to fine-tune and represent the model according to the region's heterogeneity. We will tune the values of the model parameters to obtain the best fit for our data from these sources and then assess accuracy by comparing our model predictions for independent ground truth counter measurements not used to fit the model but validated and evaluated to report accurate estimates using a different model. The primary goal of this study is to find if counter data can calibrate with user-fed Global Positioning System (GPS)-derived data from Strava to reduce ridership bias and count bicyclists frequenting the trails that are representative of all riders, both Strava users and non-Strava users. If that primary goal occurs, placing counters, collecting more accurate data based on specific characteristics, and increasing these disparate datasets' general understanding and reliability can also be improved.