Authors: Ricardo Almeida Silva; João Moura Pires; Maribel Yasmina Santos
Addresses: NOVA-LINCS Lab, Universidade Nova de Lisboa (UNL), Portugal ' NOVA-LINCS Lab, Universidade Nova de Lisboa (UNL), Portugal ' ALGORITMI Research Centre, University of Minho, Portugal
Abstract: Reasoning about spatio-temporal phenomena requires the adoption of common granularities that facilitate and enhance the comprehension of a particular phenomenon. In our day-to-day activities, spatial granules like state, province or country, and temporal granules like day, month or year, are used to index facts and to allow reasoning adopting the level of detail considered appropriate in a particular analytical context. In an era where huge amounts of spatio-temporal data are collected every day, it is crucial to model the spatio-temporal phenomena expressed in such data sets having in mind that different levels of detail can be useful in the analysis of such phenomena and that different levels of detail are related, for instance, through a spatial or temporal hierarchy. As the size and level of details of the data sets increase, the need to use multiple levels of detail that enhance our capability to achieve useful insights from data also increases. This paper presents a granularity theory devised to model spatio-temporal phenomena at different levels of detail. This granularity theory is more general than the existing granularities proposals. In fact, we relate those proposals with the presented granularity theory.
Keywords: spatio-temporal data; spatial granularity; temporal granularity; multiple levels of detail; modelling.
International Journal of Business Intelligence and Data Mining, 2015 Vol.10 No.1, pp.33 - 61
Available online: 21 Apr 2015 *Full-text access for editors Access for subscribers Purchase this article Comment on this article