Car and bike sharing are two categories of shared mobility that have presented environmental, economic and social benefits. Identification of their exogenous factors is needed to expand such concepts to new cities and to increase their performance and reliability. Additionally, ICT development has helped to increase the collection and sharing of data from the transport and geography sectors. However, the analysis and processing of such data has also become more difficult. Therefore, an automated methodology was formulated to correlate open-source arrivals and departure rates from shared transportation systems with exogenous factors from open geographic sources in multiple cities on a local scale. This methodology consisted of automated collection, analysis and processing of data as well as building of an automated model and selection of the most crucial variables using three methods: stepwise regression, GLM, and GBM. Daily average arrivals and departures in six cities in Germany (689 stations, 3.5Gb) using the bike sharing system ”Call a Bike” were used to automatically identify the relationships with exogenous factors obtained mainly from OpenStreetMap (5.9Gb). A total of 324 models were built to correlate around 200 pre-selected independent variables with the departures and arrivals from the last 3.5 years. GBM was found to fit the validation set better, whereas stepwise regression was found to perform adequately with fewer variables than other models. An indicator of the good performance of the variables selected from the resulting models were the facts that they were logical (e.g., pubs had a high influence at night) and also that they were present in the literature.     «

Car and bike sharing are two categories of shared mobility that have presented environmental, economic and social benefits. Identification of their exogenous factors is needed to expand such concepts to new cities and to increase their performance and reliability. Additionally, ICT development has helped to increase the collection and sharing of data from the transport and geography sectors. However, the analysis and processing of such data has also become more difficult. Therefore, an automated...     »