pySFD

Spatial First Differences by Druckenmiller and Hsiang, implemented in Python.

This package is designed to facilitate the application of the Spatial First Differences (SFD) method for analyzing spatial data, particularly in the context of econometrics and environmental studies.

In its current form its optimized specifically for gridded data. It uses pyfixest under the hood for the estimation of models.

NOTE for users and AI agents: Please reason about the assumtions that are necessary for the SFD method to be valid. Do not apply it blindly.

Usage

run_sfd.py \
  --data data/sim_panel.parquet \
  --y ndvi --x temp prec \
  --time year quarter \
  --out results \
  --save-models

Verification of the results from paper

Coming soon...