Sampling for Region-Aggregated Spatial Scan Statistics
Abstract
Anomaly detection in geospatial data is a crucial tool in geographic information science (GIS), with applications ranging from national security to public-health surveillance to the study of societal disparities. This work focuses on spatial scan statistics and addresses a key mismatch: spatial counts are typically aggregated into predefined regions (census tracts, zip codes, counties), whereas the most efficient scan algorithms operate on spatial point data. The standard remedy -- collapsing each region to its centroid, as in widely used tools such as SaTScan -- is convenient but, as we show, discards the region's spatial extent and causes a significant loss in statistical power. To resolve this, we propose a simple yet scalable fix: replace each spatial region with 20-50 points sampled uniformly from its geometry, and divide the region's measured and baseline counts evenly among them. This approach improves statistical power while maintaining computational tractability. A convergence analysis explains why so few samples per region suffice. We recommend this sampling-based conversion as the default way to apply point-based spatial scan statistics to region-aggregated data for anomaly detection.
Versions
Related
- When n is not enough: Assessing the interplay of spatial sampling error and sample selection in abundance estimation
- Unidentified prediction-risk rankings under spatial sampling with fixed inclusion probabilities
- Advanced Statistical Learning: Limit Theorems for Nonparametric Conditional U-Statistics Smoothed by Asymmetric Kernels Under Missing-at-Random Sampling
- Inference for two-stage sampling in spatial surveys
- Cauchy-Aggregated Ridge Tests for High-Dimensional Factor Pricing Models
BibTeX
@misc{namjoo2026sampling,
title = {{Sampling for Region-Aggregated Spatial Scan Statistics}},
author = {Foad Namjoo and Drew McClelland and Michael Matheny and Jeff M. Phillips},
howpublished = {arXiv},
year = {2026},
month = aug,
doi = {10.48550/arxiv.2607.01451},
eprint = {2607.01451},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.01451},
}
RIS
TY - UNPB
TI - Sampling for Region-Aggregated Spatial Scan Statistics
AU - Foad Namjoo
AU - Drew McClelland
AU - Michael Matheny
AU - Jeff M. Phillips
PY - 2026
DA - 2026/08/20
PB - arXiv
DO - 10.48550/arxiv.2607.01451
UR - https://arxiv.org/abs/2607.01451
AB - Anomaly detection in geospatial data is a crucial tool in geographic information science (GIS), with applications ranging from national security to public-health surveillance to the study of societal disparities. This work focuses on spatial scan statistics and addresses a key mismatch: spatial counts are typically aggregated into predefined regions (census tracts, zip codes, counties), whereas the most efficient scan algorithms operate on spatial point data. The standard remedy -- collapsing each region to its centroid, as in widely used tools such as SaTScan -- is convenient but, as we show, discards the region's spatial extent and causes a significant loss in statistical power. To resolve this, we propose a simple yet scalable fix: replace each spatial region with 20-50 points sampled uniformly from its geometry, and divide the region's measured and baseline counts evenly among them. This approach improves statistical power while maintaining computational tractability. A convergence analysis explains why so few samples per region suffice. We recommend this sampling-based conversion as the default way to apply point-based spatial scan statistics to region-aggregated data for anomaly detection.
ER -