Exploring multi-scale variability in hotspot mapping: A case study on housing prices and crime occurrences in Heerlen

Abstract

This paper examines the use of varying scaling in hotspot mapping as an exploratory spatial data analysis technique. The ambiguous geographic relationships between different crime types and housing prices are used to highlight the necessity of the multi-scale approach in effective pattern recognition. Hotspot maps were created using crime and housing price data from the city of Heerlen, The Netherlands. First, insights generated from observing the hotspot maps suggest a consistent spatial pattern of all crime types except home burglaries. By analyzing these relationships further with correlation analysis, it was found that all crime types except home burglaries have a persistent negative relationship with housing prices. The relationship between home burglaries and housing prices was found to differ in direction and significance depending on location and scale of the hotspot maps. A method for fine-tuning the scale is developed and presented. The fine-tuned results enrich hotspot mapping and highlight the need to consider scale variability properties in spatial data analysis.