Store placement (location) is one of the most important factors in offline commerce because it needs to both maximise footfall and have localised sociodemographic relevance. Computer vision can be used to analyse passing vehicular traffic to infer demographics, such as from the value of a car.

Further, we predicted the number of passers-by on a given location with specific characteristics using large scale GPS data. This quantified and defined for many different target locations in the US with an accuracy to 150m.