PySpatialStats

Spatial statistics and geographically weighted modeling in Python

pyspatialstats

Latest release: 0.3.0 · Requires: Python ≥ 3.9 · License: BSD-3-Clause

A unified Python framework for spatial statistics and geographically weighted modeling — weights and ESDA, point patterns, spatial regression, clustering, disease mapping, interpolation, sampling, and GWR.

pip install git+https://github.com/zia207/PySpatialStats.git
Package GitHub
Tutorials Getting started · Chapter 0
Author PySpatialStats Contributors

What is it?

PySpatialStats (spatialstats) is a comprehensive package for spatial data analysis:

  • Core containers, CRS helpers, and distance utilities
  • Spatial weights (contiguity, distance, graph) with libpysal interoperability
  • ESDA — Moran’s I, Geary’s C, Getis-Ord G, LISA, join counts
  • Point patterns — intensity, Clark–Evans, G/F/J, Ripley’s K/L, CSR envelopes
  • Spatial regression — OLS diagnostics, Lagrange Multiplier tests, lag, error, and Durbin models
  • Clustering — Gi* hot spots, LISA clusters, scan statistics, DBSCAN, SKATER
  • Disease mapping — SMRs, empirical Bayes, and BYM (built-in sampler or PyMC)
  • Interpolation — IDW, trend surfaces, variograms, kriging, and areal transfer
  • Sampling — random, systematic, stratified, GRTS, and cLHS designs
  • Geographically weighted models — GWR, MGWR, GW machine learning, GNNWR, and GTNNWR

Quick start

from spatialstats import Queen, Moran, GWR
from spatialstats.datasets import load_lattice, load_georgia

gdf, y_col = load_lattice(n=8)
w = Queen(gdf)
print(Moran(gdf[y_col], w, permutations=199).summary())

georgia, features, target = load_georgia()
gwr = GWR(bandwidth=2.0, fixed=True, kernel="bisquare")
gwr.fit(georgia[features].values, georgia[target].values, georgia[["x", "y"]].values)
print(gwr.summary())

Installation

pip install git+https://github.com/zia207/PySpatialStats.git

Optional extras come from the same repository:

pip install "pyspatialstats[viz] @ git+https://github.com/zia207/PySpatialStats.git"        # + matplotlib / folium
pip install "pyspatialstats[deep] @ git+https://github.com/zia207/PySpatialStats.git"       # + PyTorch (GNNWR, GTNNWR)
pip install "pyspatialstats[boosting] @ git+https://github.com/zia207/PySpatialStats.git"   # + XGBoost, LightGBM
pip install "pyspatialstats[libpysal] @ git+https://github.com/zia207/PySpatialStats.git"   # + libpysal weight converters
pip install "pyspatialstats[bayes] @ git+https://github.com/zia207/PySpatialStats.git"      # + PyMC / ArviZ
pip install "pyspatialstats[gam] @ git+https://github.com/zia207/PySpatialStats.git"        # + pyGAM
pip install "pyspatialstats[all] @ git+https://github.com/zia207/PySpatialStats.git"        # everything

From a local checkout:

git clone https://github.com/zia207/PySpatialStats.git
cd PySpatialStats
pip install -e ".[all]"

Package layout

Module Description
spatialstats.core SpatialData, SpatialResult, CRS, distance
spatialstats.weights Queen, Rook, KNN, kernel, and graph weights
spatialstats.esda Global and local autocorrelation, join counts
spatialstats.viz Choropleth, Moran scatter, LISA maps
spatialstats.pointpattern Intensity, Ripley’s K/L, CSR envelopes
spatialstats.regression OLS, LM tests, SLM, SEM, SDM, impacts
spatialstats.cluster Gi*, LISA clusters, scan statistic, SKATER
spatialstats.bayes SMR, empirical Bayes, BYM / ICAR
spatialstats.interpolate IDW, variograms, kriging, areal interpolation
spatialstats.sampling SRS, GRTS, cLHS, and design-based estimators
spatialstats.gwmodel GWR, MGWR, GWML, GNNWR, GTNNWR, GTWR
spatialstats.datasets Teaching datasets used in the tutorials

Geographically weighted models

Module Model Description
linear GWR Geographically weighted regression
linear MGWR Multiscale GWR
linear GWGLM GW generalised linear models
linear GWSS GW summary statistics
spatiotemporal GTWR Geographically and temporally weighted regression
ensemble GWRandomForest GW random forest
ensemble GWXGBoost GW XGBoost
ensemble GWLightGBM GW LightGBM
deep GNNWR Neural-network weighted regression
deep GTNNWR Spatiotemporal neural-network weighted regression
multivariate GWPCA GW principal component analysis
network NetworkGWR Network-distance GWR

Tutorials

The sidebar links a full course, from installation through a capstone. Start with Getting Started or Chapter 0.1.

Part Topic
0 Setup, data types, CRS
1 Point patterns
2 Spatial autocorrelation
3 Spatial regression
4 Cluster analysis
5 Disease mapping
6 Geographically weighted models
7 Interpolation
8 Sampling and design
9 Spatio-temporal analysis
10 Capstone

References

  • Anselin, L. (1995). Local indicators of spatial association — LISA. Geographical Analysis, 27(2), 93–115.
  • Besag, J., York, J., & Mollié, A. (1991). Bayesian image restoration, with two applications in spatial statistics. Annals of the Institute of Statistical Mathematics, 43, 1–20.
  • Fotheringham, A. S., Brunsdon, C., & Charlton, M. (2002). Geographically Weighted Regression. Wiley.
  • Oshan, T. M. et al. (2019). mgwr: A Python implementation of multiscale geographically weighted regression. ISPRS IJGI, 8(6), 269.
  • Du, Z. et al. (2020). Geographically neural network weighted regression. IJGIS, 34(7), 1353–1377.
  • Wu, S. et al. (2021). Geographically and temporally neural network weighted regression. IJGIS, 35(3), 582–608.
Back to top