Understanding urban land use from digital footprints of crowds
The established view on semantic organization of space is based on the concept of “land use”, which corresponds to an aggregate perspective on the use of an area (e.g. agriculture, residential, business, etc.). The characterization of urban block is built on the human activities that happen there, however a more disaggregated and dynamic view is now possible due to availability of new techniques and technologies. In fact, this should become a more natural way to profile the places.
Understanding population dynamics by type, neighborhood, or region would enable customized services (and advertising), as well as the accurate timing of urban service provisions, such as scheduling transit service based on daily, weekly, or monthly mobility demand. In general, more synchronous management of service infrastructures clearly could play an important role in urban mobility management. Traditionally, urban planning relies on census survey conducted every 5-10 years and has shortcomings both in terms of spatial and temporal scale. The wide deployment of pervasive computing devices (cell phone, smart card, GPS devices and digital cameras) provide unprecedented digital footprints, telling where and when people are. In former projects, we developed a methodology for detecting the presence and movement of crowds through their digital traces (flickr photo, cell phone logs, smart card record and taxi/bus GPS traces).
This fine grained analysis, up to the level of the establishment, makes a big leap in terms of understanding the use of space for the purposes of urban planning and management. In recent work, we have presented several perspectives on extracting semantics of the place from online information. A further step shapes on the intersection of such generic information about space with other digital footprints, such as cell phone usage or taxi demand. An essential scientific contribution of this proposal will be on development of new techniques for land use analysis supported on semantic enriched POIs.