geoRDDprep 0.1.1

Preparing data for regression discontinuity design

geoRDDprep

geoRDDprep is a Python package designed to streamline the data preparation process for Geographical Regression Discontinuity Design (GeoRDD). It provides efficient tools for spatial joins, polygon-to-line conversions, and implementing the Turner et al. (2014) algorithm for assigning points to boundaries.

Features

• points_in_polygon: Efficiently assign points to polygons (e.g., addresses to school districts).
• turner: Assign points to LineStrings based on orthogonal distance criteria (Turner et al., 2014).
• poly_to_line: Convert Polygon geometries to LineStrings for boundary analysis.
• drop_tiny_lines: Filter out small, noisy line segments to improve analysis quality.
• remove_sliver: Clean up sliver polygons using Voronoi diagrams.
• remove_overlaps: Remove overlapping segments between line datasets.

Installation

You can install the package directly from the source:

pip install .

Or, if you are developing:

pip install -e .

Usage

1. Assign Points to Polygons

Map addresses or other points to their respective administrative regions.

import geopandas as gpd
from geoRDDprep import points_in_polygon

# Load your data
points = gpd.read_file("addresses.geojson")
districts = gpd.read_file("districts.geojson")

# Assign points to districts
# The resulting GeoDataFrame will have columns from 'districts' suffixed with '_district'
result = points_in_polygon(points, districts, suffix_name="_district")

2. Prepare Boundaries (Polygons to Lines)

Convert polygon boundaries into lines for distance analysis.

from geoRDDprep import poly_to_line, drop_tiny_lines

# Convert polygons to lines
lines = poly_to_line(districts)

# Clean up noise by dropping very short lines (e.g., < 500 meters)
clean_lines = drop_tiny_lines(lines, method='length', meters=500)

3. Turner Algorithm

Assign points to boundaries based on distance and orthogonality.

from geoRDDprep import turner

# Match points to the nearest boundary within 15 meters
# 'turner_pass' column will be True if the point satisfies the criteria
matched_data = turner(points, clean_lines, orth_distance=15)

Requirements

• geopandas
• shapely
• numpy
• pandas
• scipy

License

MIT License