Analysis of spatial data refers to computer-based analytic approaches that study entities using their topological, geometric, or geographic properties. Spatial data have recently become popular due to web-applications like Google Earth, which handle the locations of entities on digital maps. Moreover, spatial data and its analysis are widely used in various other fields, such as astronomy, geographical information science (GIS), VLSI design (chip fabrication engineering that builds complex wiring structures), biology, epidemiology, sociology, demography, statistics, remote sensing, computer science, and scientific modeling. In this course, we will focus on issues related to: (a) the management of spatial data and (b) their analysis using spatial data mining techniques. Spatial data management involves studying spatial databases methods that optimize storage and querying data related to objects in space (points, lines and polygons). Spatial databases differ from relational databases, because the latter can only understand numeric and character types of data, thus additional functionality needs to be added for databases to process spatial data types. We are going to revisit design methodologies (ER diagrams for spatial data), physical storage and indexing issues (spatial indexing), querying and its optimization (SQL for spatial data).
Moreover, the spatial data mining methods we will consider include spatial classification, clustering, and spatial association rule mining. All these methods differ from their corresponding, conventional data mining methods, because they have to consider several characteristics of spatial data, such as auto-correlation and higher-dimensionality. Focus will also be given on obtaining hands-on experience using a real spatial-database (PostgreSQL) and implementing all theoretical concepts in its framework.