A traffic flow map, i.e. a map of the traffic flow levels on the road segments in a road network, is a fundamental quantity in transport planning. Whilst traffic flow maps are a rich source of information about vehicle mobility patterns, they are also costly in terms of time and resources to compute for any reasonably sized road network.

To alleviate this cost burden, we can use empirical trajectory data, aka Floating Car Data (FCD), as our input data source since they can be acquired with low marginal cost whilst offering extensive spatio-temporal coverage. Click on each city name in the header banner for interactive flow maps computed from trajectory data from that city.

In a preview of these flow maps in the image comparison below, we zoom into 376 GPS trajectories collected in Hannover, Germany. The green circles of the GPS trajectories can deviate from the road network (measurement errors), and the vehicle location between the recorded GPS points is unknown (sampling errors).

If we move the slider from right to left, the error-prone GPS empirical trajectories are replaced with single flow lines, each with a single traffic flow value. Our innovative flow map visualises intuitively and unambiguously the traffic flow on each road segment.

Details of this innovation based on aligning road segments to the most locally representative segment can be found in the ArXiv preprint arxiv:2406.17500. Any questions about the methodology can be addressed to Tarn Duong tarn(dot)duong(at)gmail(dot)com.