Tracks vs. Counters: Towards a Systematic Analysis of Spatiotemporal Factors Influencing Correlation

Mobility data from tracking apps and stationary counters are often limited by biased sampling, uneven spatial distribution and low spatial coverage, respectively. Data fusion approaches attempt to combine the advantages of both sources. However, the observed correlation of tracks and counts is often mediocre. The reasons for these differences are often ascribed to sampling bias. However, we argue that this is an oversimplification of the real relationship. We present our work in progress concept for relating tracking data and stationary counting data to critically reflect on the factors influencing their correlation and how this can inform data fusion approaches.