1 Projection of road sensors to the Dutch road network Martijn Tennekes (m.tennekes@cbs.nl) 1 and Marco Puts (m.puts@cbs.nl) 1 Keywords: Visual inspection, spatial data, big data, traffic statistics 1. INTRODUCTION Road sensors measure the number of passing vehicles every minute. In the Dutch network of highways, there are approximately 20 thousand of those road sensors, resulting in a huge data source. A key step in the production of traffic statistics from this data source is to project the geographic locations of the road sensors on the Dutch road network. To achieve this, road segments have to be defined based on the locations of the road sensors and subsequently the lengths of these segments. The method that is described in this paper consists of two main parts. First, the main routes per highway per direction are deduced from the detailed road network. Second, these main routes split into road segments based on the locations of the road sensors, and the entrance and exit ramps. We illustrate that visual inspection throughout the whole process is crucial. All geographic data inspection and editing is done in R, especially with the recently developed package tmap [1]. This package contains a flexible plotting method that is similar to ggplot2 [2], but tailored to spatial data. Also, some processing functions that were needed for this project were added to the package. 2. METHOD The first part of the process is the necessary pre-processing of geographic locations of the road sensors and the road network which is described in section 2.1. The extraction of the main routes is discussed in section 2.2. Finally, in section 2.3, we describe the calculation of the road segment lengths. 2.1. Pre-processing geographic location data For statistical interference of geographic data, it is important to use a proper map projection. Unprojected map coordinates, known as latitude-longitude coordinates, often lead to inaccurate measurements for distances, area sizes, and directions. A good map projection preserves one or more of these properties. For the task at hand we use the Dutch National Grid (Rijksdriehoekstelsel), which is a Cartesian coordinate system that is optimized for the Netherlands. It preserves distances, which means that the difference between any two coordinates in the Netherlands corresponds approximately to the real distance in meters. Besides the geographic coordinates, the metadata of the road sensors consist of the road names, the direction, and the type of carriageway. Only the road sensors from the main carriageways of the main Dutch highways were selected for the further process. 1 Statistics Netherlands, Heerlen, The Netherlands