spatial-social-analysis

Compute Moran's I to detect spatial autocorrelation by constructing spatial weights and calculating global clustering statistics. This identifies whether nearby observations influence each other, revealing geographic dependence across your dataset.

Fit a spatial lag model (SAR) to handle spatial spillovers by estimating spatial dependence parameters. This models how outcomes in neighboring regions influence your target variable, correcting bias from geographic clustering.

LISA, or Local Moran's I, is used in hotspot detection to identify local clusters of similar values. It pinpoints specific hotspots and coldspots, enabling detailed choropleth mapping of significant spatial patterns.

Yes, you can use geopandas with spreg to run spatial econometrics including spatial regimes. This estimates spatial error models and spatial Durbin models while accounting for structural shifts across different geographic boundaries.

Prevent spatial data leakage in machine learning validation by using spatial cross-validation with coordinate-based grouping. This ensures geographically connected observations stay in the same fold, producing reliable spatial ML performance metrics.

Yes, this skill supports geographic boundary designs and accessibility measurement. It quantifies spatial accessibility using 2SFCA methods and runs boundary-based geographic RDD workflows for analyzing service access across regions.