SPATIAL ANALYSIS OF EXTERNAL INFLUENCES ON TRAFFIC ACCIDENTS USING OPEN DATA
Keywords: traffic accidents, external influences, environmental context, open data, machine learning
Abstract. In this article traffic accidents in Germany are investigated based on different environmental impacting factors. These factors are related to the accident surrounding like road characteristics, weather information, population density as well as POIs. For this purpose different open data source are used, like OpenStreetMap and an open weather API. These input data sources are processed in order to generate different categories of features, describing the on-site situation of an accident. Using the k-means clustering, six clusters were identified. In a consecutive investigation, each cluster is not only described with respect to the feature space but also in regards to their spatial distribution. Based on these findings, a regional comparison is done across the state boarders of Germany.