Geospatial Analysis in R
This page hosts two tutorials on working with spatial data in R for political science research.
Quasi-Experimental Designs Using Spatial Data
This tutorial, on spatial data for causal identification, is based on a workshop I prepared to teach at the 2026 edition of the Summer School of Political Methodology at the University of Mannheim.
You will learn:
- Why spatial autocorrelation matters for causal inference (SUTVA violations, spillovers, and spatial confounding)
- Geographic regression discontinuity designs
- Difference-in-differences with spatially assigned treatment
- Matching on spatial covariates
- Using spatial regression (SAR / SEM) as a robustness diagnostic
This tutorial assumes familiarity with the basics of R and causal inference. The GIS Visualization tutorial below is a useful prerequisite for the spatial-data mechanics.
GIS Visualization and Analysis in R
I designed this tutorial originally for the MIT Political Science Methods Workshop series for graduate students and faculty members. It teaches how to make maps and geospatial calculations in R.
You will learn:
- What are the main components of spatial data
- How we can combine spatial data with other data sources
- How we can create and manipulate maps
- Spatial regression techniques
This tutorial is good for students with some familiarity with R, but does not require previous knowledge of geospatial analysis. The updated version uses the modern sf / terra stack.