An Introduction to Spatial Data Mining

Date of Submission: 
August 8, 2018
Report Number: 
Report PDF: 

The goal of spatial data mining is to discover potentially useful, interesting, and non-trivial patterns from spatial datasets. Spatial data mining is important for societal applications in public health, public safety, agriculture, environmental science, climate etc. For example,in epidemiology, spatial data mining helps to find areas with a high concentrations of disease incidents to manage disease outbreaks. Computerized methods are needed to discover spatial patterns since the volume and velocity of spatial data exceeds the number of human experts available to analyze it. In addition, spatial data has unique characteristics like spatial autocorrelation and spatial heterogeneity which violate the i.i.d (Independent and Identically Distributed data samples) assumption of traditional statistics and data mining methods. So, using traditional methods may miss patterns or may yield spurious patterns which are costly (e.g., stigmatization) in spatial applications. Also, there are other intrinsic challenges such as MAUP (Modifiable Areal Unit Problem) as illustrated by a current court case debating gerrymandering in elections. Spatial data mining considers the unique characteristics, and challenges of spatial data and domain knowledge of the target application to discover more accurate and interesting patterns.In this article, we discuss tools and computational methods of spatial data mining, focusing on the primary spatial pattern families: hotspot detection, colocation detection, spatial prediction and spatial outlier detection. Hotspot detection methods use domain information to model accurately more active and high density areas. Colocation detection methods find objects whose instances are in proximity of each other in a location. Spatial prediction approaches explicitly model neighborhood relationship of locations to predict target variables from input features. The goal of spatial outlier detection methods is to find data that are different from their neighbors.