prantogg opened a new pull request, #5:
URL: https://github.com/apache/sedona-spatialbench/pull/5
This PR introduces two major updates to the spatial data generator:
1. Continent-Bounded Affines
- Replaces the single full-world affine with continent-bounded affine
transforms (Africa, Europe, South Asia, North Asia, Oceania, South America,
South North America, North North America).
- This improves realism by distributing geometries only over land
masses, introducing natural skew land masses in the coverage.
- The affine is now fixed and not user-configurable, ensuring consistent
and reproducible bounding across runs.
2. New Distributions: Thomas & Hierarchical Thomas
- Thomas Distribution (Neyman–Scott process)
- Adds parent clusters (hotspots) with offspring distributed via
Gaussian spread.
- Cluster sizes follow a Pareto distribution, introducing realistic
skew where some hotspots dominate.
- Suitable for mimicking real-world trip data (urban hotspots, uneven
density).
- Hierarchical Thomas Distribution
- Extends Thomas by introducing a two-level hierarchy:
- Cities (top-level hotspots)
- Subclusters (neighborhood-level clusters within cities)
- City and subcluster sizes both follow Pareto distributions,
capturing heavy-tailed skew at multiple levels.
- Configurable spreads (sigma_city, sigma_sub) let users tune how
dispersed or tight clusters are.
- Ideal for modeling realistic urban structures where trips/buildings
concentrate into cities → neighborhoods → hotspots.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
