Is it possible to make reliable predictions of the number of people who walk or drive down a particular street in Germany? Even without conducting special traffic surveys? Yes, and the solution lies in spatial data mining of different sources.
This is good news, especially for marketing strategists, traffic and logistics specialists, and billboard advertisers: They can now choose the best site for their cash dispensers, branch offices, retail outlets or advertising media. All thanks to the technique of spatial data mining, as developed and put into practice by researchers at the Fraunhofer Institute for Autonomous Intelligent Systems AIS. It involves the use of a set of calculation methods that allow large volumes of data, such as the customer databases held by major companies, to be systematically searched and filtered according to geographical criteria.
One of the many conceivable applications of spatial data mining is the frequency atlas produced by an AIS research team on behalf of the German outdoor advertising association (Fachverband fόr Aussenwerbung FAW). It provides an estimation of the number of people who can be expected to walk or drive along certain stretches of road in Germany, at different times of the day, even in places where traffic surveys have never been carried out. The database was originally compiled as a means of determining fair prices for different billboard locations. After all, advertising agencies work on the principle that the greater the exposure of a particular site, the higher the prices charged. All advertisers want their posters to be seen as often as possible by people belonging to their target group. At present, the frequency atlas covers all 82 major cities in Germany with a population of over 100,000 inhabitants. The next stage of the project is to extend the database to cover all cities with a population of 50,000 or more.
A variety of sources are used to provide the data for analysis. The GfK Group market research institute, for example, supplies empirical data gathered during traffic surveys. These data are evaluated together with the results of socio-demographic studies conducted by various providers of geographical data services, including information on points of interest such as railroad stations, gas stations and restaurants. And road networks for which there is a routing management system provide information on the directions in which the majority of pedestrian or vehicular traffic flows on various roads. “We have combined all of these datasets in a spatially oriented model that provides a high level of accuracy,” relates Michael May, head of the Knowledge Discovery department at the AIS. “The method we use to combine spatial data mining and geographical information systems is unique of its kind in the world.”
Cite This Page: