GPS probe data for Traffic Analytics

The GPS probe data used in Traffic Analytics comes from a mixed fleet of 170+ providers around the world, each of which specializes in certain regions and countries. The mix reflects a broad sample of the vehicles on the road, though some pedestrian, bicycle, and transit data is unavoidably present. The data can come from connected vehicles, the most important and accurate source, navigation systems, fleet telematics systems, or mobile devices. GPS probe data is usually treated as a set of instantaneous speed observations, which are subject to natural sampling variations, particularly on arterial roads.

The Traffic Analytics probe data isn't modeled. Gaps exist, and you might include statistical outliers, such as in level-of-service studies. HERE cleans and normalizes its probe data to assure the highest quality, using advanced algorithms specifically tuned for historical data, and more computationally intense than would be practical in real-time traffic feeds.

Probe speed data is averaged in five-minute increments (or optionally, 15, and 60-minute increments), and presented with other measures such as minimum, maximum, standard deviation, percentiles, and a confidence measure. For convenience, a reference “free flow” speed is also included in the dataset, representing the speed at which traffic flows when not congested. This helps in road network performance analysis, which derives how often observed speed falls significantly below the free flow or not congested speed. Traffic Analytics enhances with harmonic means, vehicle distance traveled, and probe counts.

The Traffic Analytics portal and API also helps manage the challenge of large data volumes typical of probe-based analysis with filtering options, while still supporting multi-gigabyte datasets, whenever you need them.