In addition to this geographic data, network operators have ­access to the above described service and user specific information that are typically captured as market cell data with a high spatial resolution; e.g. 3G microcells in downtown areas have a radius of less than one hundred meters.

The information extracted from user- and network-specific data can be used, for example, to identify geographic and ­radio ­related causes for poor service quality, for the automated ­estimation of the traffic demand based on geographic, demographic, and radio related information, or even for the customer segmentation.

Already simple studies demonstrate that the observable ­variation of voice traffic in a radio cell over the day generally correlates ­closely with the spatial location of the cell. This is easy to ­explain, because the use of voice services is determined by human ­behavior: people are at home, at work, moving around, shopping. This results in characteristic service usage profiles which can be used to easily identify the dominant human in a ­particular cell. Cells with a high level of road traffic typically have peaks of the voice traffic during the rush hour. Cells which cover residential areas typically show increasing voice traffic after the end of office hours. By means of simple algorithms that compare selected clusters according to various criteria it is possible to identify the dominant user behavior of all cells in a mobile network.

Analogous analysis extend the level of detail to user groups. For example, the cells in which certain user groups work and live can be identified because users regularly spend a large amount of time and use mobile services in these cells.

If all of the available data sources are systematically correlated, it is finally possible to draw quite a precise picture of the customer behavior. Network operators therefore can understand what ­customer groups use what type of services at what locations and can therefore provide their services aligned to the real demand.

The knowledge of the functional dependency of customer ­characteristics on spatial factors then helps to understand the driving forces of the service usage and how they can be ­supported by marketing instruments. In this way it can be estimated what customer segments are using what services at which locations during what activities – assuming that services are generally only used in certain situations which are highly dependent on the current location of the customer.