Spatial Analysis on Value-Based Quadtrees of Rasterized Vector Data

Abstract

Mobility data science offers insights into the complex interconnections of spatial data of moving objects and their surroundings, often based on a combination of vector and raster data. For example, mobility traces are usually in vector format, weather data are often in raster format. Yet, available spatial analysis tools for exploratory data science push data scientists towards one or the other, providing only limited support for the respective other. In this paper, we contribute to this problem space with a value-based quadtree index, which serves as a bridge builder to support joint spatial analysis on vector and raster data leveraging their unique autocorrelation property. We achieve a 90% reduction in median Point-in-Polygon query latency, while keeping the accuracy of query responses at equal level.

Figures (6)

• Figure 1: We rasterize the trajectory vector data (left) with a resolution of 5$m^{2}$ to a raster representation (middle) and create a quadtree index (right).
• Figure 2: The first dataset containing trajectories (a) is rasterized to a 5$m^{2}$ raster (b). The second dataset is polygon-based and rasterized to the same raster structure ensuring that both are aligned. Then, a quadtree index is created on both rasters (c).
• Figure 3: Intersecting the datasets allows running queries faster by using the speedup provided by the quadtree representations.
• Figure 4: query duration on three data formats vector, raster, and quadtree sorted by resolution of data with high spatial autocorrelation (parks in Berlin from ). As the resolution can be modified for raster and quadtree data only, we compare different resolutions to the original vector data, indicated as "5".
• Figure 5: query duration on the three data formats vector, raster, and quadtree sorted by resolution of data with low spacial autocorrelation (cycling trajectories in Berlin from SimRa).