Understanding and Visualizing the District of Columbia Capital Bikeshare System Using Data Analysis for Balancing Purposes

Bike sharing systems' popularity has consistently been rising during the past years. Managing and maintaining these emerging systems are indispensable parts of these systems. Visualizing the current operations can assist in getting a better grasp on the performance of the system. In this paper, a data mining approach is used to identify and visualize some important factors related to bike-share operations and management. To consolidate the data, we cluster stations that have a similar pickup and drop-off profiles during weekdays and weekends. We provide the temporal profile of the center of each cluster which can be used as a simple and practical approach for approximating the number of pickups and drop-offs of the stations. We also define two indices based on stations' shortages and surpluses that reflect the degree of balancing aid a station needs. These indices can help stakeholders improve the quality of the bike-share user experience in at-least two ways. It can act as a complement to balancing optimization efforts, and it can identify stations that need expansion. We mine the District of Columbia's regional bike-share data and discuss the findings of this data set. We examine the bike-share system during different quarters of the year and during both peak and non-peak hours. Findings reflect that on weekdays most of the pickups and drop-offs happen during the morning and evening peaks whereas on weekends pickups and drop-offs are spread out throughout the day. We also show that throughout the day, more than 40% of the stations are relatively self-balanced. Not worrying about these stations during ordinary days can allow the balancing efforts to focus on a fewer stations and therefore potentially improve the efficiency of the balancing optimization models.