Visualize the Invisible & Foresee the Unforeseen
Sensors are widely used on streets to detect and evaluate traffic condition. For instance, in-ground and other local sensors can count numbers of all passing-by bicycles with a high temporal resolution. However, in-ground sensors can not process well with a high spatial resolution, unless sensors are equipped on every street corner. Besides, some data source such as Stravo Metro is questioned by public due to data bias. Therefore, are there any other approaches that allow us to analyze bicycle data more precisely?
Daniel I. Patterson gives the answer from a new perspective. Instead of obtaining data from sensors, he works on crowdsourcing data to gain insight into cyclists behavior when they use the local bicycle system. Based on start-and-end points data, he speculates and predicts route preferences of bicycle users at different time of the day, and visualizes results by using GraphHopper API and Stplanr package. This source data package comes from ‘Minneapolis — St.Paul Nice Ride’ bicycle sharing program which is currently operated by Motivate.
Although he uses Twin Cities data to do research on its bicycle system, issues are similar in other cities with bicycle sharing program including New York, Washington D.C, Portland, Houston and so on. This method can also be widely used for other public bicycle management, such as optimization of bicycle lanes, expansion of bicycle system, road planning.
