Merge pull request #15 from PolicyEngine/paper

congressional district tesselation

Author: baogorek

reweighting/cd-plotting-tiles.py

@@ -0,0 +1,81 @@
+import geopandas as gpd
+from shapely import affinity
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+
+
+# https://docs.google.com/spreadsheets/d/13XkF59JKzvw4SeSq5mbgIFrJfYjK4amg9JoQE5e9grQ/edit?gid=0#gid=0
+# https://www.dailykos.com/stories/2019/7/23/1873464/-We-took-the-best-map-ever-of-US-congressional-districts-and-made-it-even-better
+# https://www.primarycast.us/house-races/KY_2
+
+# ----------------------------------------------------
+# read the hex-cartogram shapefile you downloaded
+# ----------------------------------------------------
+shp_path = r"/mnt/c/devl/data/pe/HexCDv31/HexCDv31.shp"
+
+gdf = gpd.read_file(shp_path)
+
+# QUICK sanity-check: take a look at the attribute names
+print(gdf.columns.tolist())
+print(gdf.head(3))
+
+# ----------------------------------------------------
+nc = gdf[gdf["STATEAB"] == "NC"]
+
+# ----------------------------------------------------
+#  add your 14 numbers
+# ----------------------------------------------------
+my_numbers = np.random.rand(14)
+
+vals = pd.DataFrame({
+    "CDLABEL": range(1, 15),
+    "metric":  my_numbers
+})
+
+nc["CDLABEL"] = nc["CDLABEL"].astype(int)
+
+# ------------------------------------------------------------------
+#  merge and plot
+# ------------------------------------------------------------------
+nc_plot = nc.merge(vals, on="CDLABEL", how="left")
+
+fig, ax = plt.subplots(figsize=(6, 7))
+
+# draw boundaries so the hexes have borders
+nc_plot.boundary.plot(ax=ax, edgecolor="black", linewidth=0.4)
+
+# fill by your metric
+nc_plot.plot(column="metric",
+             cmap="viridis",        # pick any Matplotlib colormap
+             linewidth=0,
+             legend=True,
+             ax=ax)
+
+ax.set_title("North Carolina congressional districts", fontsize=13)
+ax.set_axis_off()
+plt.tight_layout()
+plt.show()
+
+# Now rotate to make the state flat
+# pick a single pivot
+pivot = nc_plot.unary_union.centroid    # nc_plot is your merged GeoDataFrame
+
+# rotate every geometry around *that* pivot
+angle = -12
+nc_rot = nc_plot.copy()
+nc_rot["geometry"] = nc_rot.geometry.apply(
+    lambda g: affinity.rotate(g, angle, origin=pivot)
+)
+
+# 3) plot
+fig, ax = plt.subplots(figsize=(6, 7))
+nc_rot.boundary.plot(ax=ax, edgecolor="black", linewidth=0.4)
+nc_rot.plot(column="metric", cmap="viridis", linewidth=0, legend=True, ax=ax)
+
+ax.set_title("One Household's Weight Values Across NC's 14 Districts", fontsize=13)
+ax.set_axis_off()
+plt.tight_layout()
+plt.show()
+  
+