A Fast, Versatile, and User-Friendly Plugin for Kernel Density Analysis

Kernel density analysis (KDA) is a commonly used tool in various application domains. Some representative examples include disease outbreak detection and crime/traffic/traffic accident hotspot detection. However, KDA is a slow operation, which cannot scale to large resolution sizes and large-scale datasets. In addition, domain experts also adopt the more complicated variants of KDA, e.g., spatiotemporal kernel density analysis (STKDA) and network kernel density analysis (NKDA), which deteriorates the efficiency issues for analysis. Furthermore, existing software packages cannot simultaneously support KDA, STKDA, and NKDA and some of them require the prior programming skills, which may not be equipped by domain experts. In order to address the above efficiency, versatility, and usability issues, we develop the QGIS plugin, called Fast Density Analysis, which is based on our complexity-reduced algorithms for supporting KDA, STKDA, and NKDA. With the concise user interface, users do not need to have any prior programming skills for using this plugin. In this demonstration, we offer three location datasets (with up to 8.32 million data points) for participants to (1) compare the efficiency between this plugin and existing software packages and (2) conduct case studies to discover hotspots/hidden patterns in these datasets.