An Approach for spatial and temporal data analysis: application for mobility modeling of workers in Luxembourg and its bordering areas

In this paper, we propose a general visual analytic approach to synthesis very large spatial data and discover interesting knowledge and unknown patterns from complex data based on Origin-Destination (OD) matrices. The research studies of Tobler constitute a good basis in this topic. This paper is interested in the proposal of 2 methods entitled respectively ?Weighted Linear Directional Mean: WLDM? and ?DS-WLDM?. The latter incorporates the Dempster-Shafer theory of evidence with WLDM. Both of the developed methods are an extension of ?Linear Directional Mean: LDM? for mobility modeling. With classical techniques such as LDM among others, the results of data mapping are not intelligible and easy to interpret. However with both WLDM and DS-WLDM methods it is easy to discover knowledge without losing a lot of information which is one of the interests of this paper. This proposal is generic and it intends to be applied for data mapping such as for geographical presentation of social and demographic information (e.g. mobility of people, goods and information) according to multiple spatial scales (e.g. locality, district, municipality). It could be applied also in transportation field (e.g. traffic flow). For the application, administrative data is used in order to evaluate spatial and temporal aspects of the daily and the residential mobility of workers in Luxembourg and its bordering areas.