Explaining Spatio-Temporal Dynamics in Carsharing Twenty-second Americas Conference on Information Systems, San Diego, 2016 1 Explaining Spatio-Temporal Dynamics in Carsharing: A Case Study of Amsterdam Emergent Research Forum Paper Konstantin Klemmer University of Freiburg konstantin.klemmer@is.uni- freiburg.de Christoph Willing University of Freiburg christoph.willing@is.uni-freiburg.de Sebastian Wagner University of Freiburg & Geospin GmbH swagner@geospin.de Tobias Brandt University of Freiburg tobias.brandt@is.uni-freiburg.de Abstract We investigate customer mobility behavior by examining free-floating carsharing demand dynamics. For this purpose, we analyze rental data of a major carsharing provider in the city of Amsterdam in combination with points of interest (POIs). Connecting POI data to carsharing trips and stratifying the data along 6-hour intervals allows us to illustrate the spatio-temporal dimensions of carsharing usage, i.e. how carsharing demand changes over time and how it shifts spatially within the provider’s business area. We cluster the point data using kernel density estimation and apply a generalized linear model with Gamma distributed values on the sampled data. Our results indicate that, depending on the hour of the day, different POI categories have different, yet significant, impact on trip destinations. Our insights advance the understanding of when and for what purpose customers use carsharing, enabling providers to predict demand in existing and new business areas. Keywords Carsharing, Spatial Analytics, Location-based Services, Visualization Introduction During the last decade, carsharing has emerged as one of the most prominent examples of the sharing economy. In urban areas with a well-developed carsharing infrastructure, shared vehicles can be even more attractive than a personal car. By using carsharing, customers can save taxes and maintenance costs, while still being able to access vehicles whenever needed. The elimination of the fixed costs of car ownership also enables customers who previously could not afford these expenses to enjoy the benefits of private car travel (Martin et al. 2010). Furthermore, the increasing adoption of carsharing results in a reduction of CO2 emissions, which contributes to solving the pollution problem of many metropolitan areas (Firnkorn and Müller 2011). In addition, several carsharing companies offer, either partly or exclusively, hybrid and electric vehicles and thus contribute to environmental sustainability even further. Today’s carsharing systems face complex challenges. An advanced method of carsharing is the free-floating approach: In this model, vehicles may be driven to any location within a predefined operational area. Compared to a model with fixed parking spots or stations, free-floating carsharing offers customers more flexibility and enables new transportation opportunities such as one-way trips or the combination of carsharing and public transport. However, the problem of vehicle relocation remains unresolved, since within the operating area there are regions of relatively high demand (underflow region) and relatively low demand (overflow region), which exhibit fewer and more vehicles than needed, respectively (Lee and Park 2014). The essential step in resolving this issue is to figure out why these regions emerge in the first place. Therefore, the determinants of vehicle distribution need to be identified and the behavior of carsharing users has to be analyzed, which is why in this paper we aim to answer the following questions: brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by AIS Electronic Library (AISeL)