Open source in accessibility analyses : Modeling of street petrol station accessibility in Germany

The paper relates two closely intertwined story lines. A socioeconomic one concentrating on street petrol station accessibility in Germany’s rural areas and a geospatial one concentrating on the usability of the Open Street Map within rural studies. This is attributed to the fact that the paper builds on findings from applied research within rural studies and is intended to serve as a "practical experiences report" assessing the usability of open source GIS/-data in rural studies. Here we analysed the accessibility of street petrol stations as one core service of general interest important for the overall individual mobility of the population, especially in rural areas, based on an exemplary raster-based GIS accessibility analysis. This analysis builds upon an open source approach using PostgreSQL/PostGIS as well as the Dijkstra shortest path algorithm implemented in the Perl-module "Graph-0.94". Besides acquiring objective data on street petrol stations accessibility for policy advice, we were also interested in reviewing the usability of OpenStreetMap (OSM) data compared to commercial routing networks (ESM). Altogether the findings suggest that in Germany, street petrol stations are for the majority of the population (ESM: 99.5%/ OSM: 99.4%) quite accessible. On average the distance to the next street petrol station amounts 5.4 km (ESM)/ 5.5 km (OSM). Regions with disadvantageous accessibility are predominantly sparsely populated. The comparison of the accessibility values calculated based on ESM and OSM showed that great differences exist in a per cell comparison as well as on the community level, whereas aggregated average accessibility values for greater aggregates like counties proved to be comparable to each other. Against the background of OSM’s lower level of completeness in rural areas an interesting but unexpected result is the fact that the accessibility values differ within urban as well as rural areas within the same range. Nevertheless, considering the identiidentified shortcomings of the OSM, the data set still seems to show a lower performance than commercial data sets.