def _combine_data(
*,
location_data: gpd.GeoDataFrame,
population_data: pd.DataFrame=None,
location_index='NUMBER',
**kwargs
) -> gpd.GeoDataFrame:
location_data = utils.split_plots(location_data, location_index)
if isinstance(population_data, pd.DataFrame):
location_data = location_data.set_index(location_index)
location_data = location_data.join(population_data, **kwargs)
return location_data
def join_viewpoints_to_split_islands(
islands_df: GeoDataFrame,
viewpoints: GeoDataFrame,
split_islands: GeoDataFrame,
output_path: str,
overwrite: bool = False
) -> GeoDataFrame:
if output_path and exists(output_path) and not overwrite:
return gpd.read_file(output_path)
else:
jvp = viewpoints.join(split_islands[['FID_island']], on='FID', how='inner')
jvp = jvp[['FID_island', 'Z', 'geometry']]
ivp = jvp.join(islands_df, on='FID_island', rsuffix='_island')[
['FID_island', 'Z', 'geometry']]
ivp.to_file(output_path)
return ivp
"winddirection_80m",
"winddirection_100m",
"winddirection_200m",
]
gdf = GeoDataFrame().from_file("vector/h5_grid.shp")
for dataset in datasets:
logger.debug("Iterating over our target WTK datasets and flush the" +
"attributed results to disc as we go")
if os.path.exists("vector/" + dataset + ".shp"):
print("Existing file found (skipping): " + "vector/" + dataset +
".shp")
else:
result = attribute_gdf_w_dataset(
gdf,
hour_interval=_HOURS_PER_MONTH,
n_bootstrap_replicates=30,
dataset=dataset,
)
result.columns = [
dataset + "_" + str(n) for n in list(result.columns)
]
result = gdf.join(result)
result.to_file("vector/" + dataset + ".geojson", driver="GeoJSON")
del result
def visualize_building_counts(dg: gpd.GeoDataFrame, osm: gpd.GeoDataFrame):
# get n range
max_n = max(dg.num_buildings.max(), osm.num_buildings.max())
min_n = 0
# visualize num_buildings distribution
fig, ax = plt.subplots(1, 2, sharey='row', sharex='col')
bins = range(0, max_n + 1, 5000)
dg.num_buildings.hist( ax = ax[0], bins = bins)
osm.num_buildings.hist(ax = ax[1], bins = bins)
ax[0].set_title("DigitalGlobe")
ax[1].set_title("OpenStreetMap")
fig.suptitle("Histogram of Per-Block Building Counts for Freetown, SL")
plt.semilogy()
# plt.savefig("n_histogram_sl.png", dpi=300, bbox_inches="tight")
plt.show()
# num_buildings choropleth
fig, ax = plt.subplots(1, 2, sharey='row', sharex='col')
plot_opts = {
"column" : "num_buildings",
"cmap" : "PiYG",
# "scheme" : "natural_breaks",
"scheme" : "User_Defined",
"edgecolor": "grey",
"linewidth": 0.1,
"legend" : True,
"categorical" : False,
"markersize" : 0,
"classification_kwds" : {"bins" : [4270, 16922, 47521, 122413, 160779, max_n]} # numbers taken from natural_breaks
}
# dummy polygons to make choropleth color scheme consistent
dgd = gpd.GeoDataFrame(pd.concat([pd.DataFrame([
["d0", shapely.geometry.MultiPoint(osm.centroid).centroid, 0, [], 35],
], columns = dg.columns), dg.copy()]).reset_index(), geometry = "block_geom")
dgd.plot(ax=ax[0], **plot_opts)
ax[0].get_xaxis().set_visible(False)
ax[0].axes.get_yaxis().set_visible(False)
# osmd.plot(ax=ax[1], **plot_opts)
osm.plot(ax=ax[1], **plot_opts)
ax[1].get_xaxis().set_visible(False)
ax[1].axes.get_yaxis().set_visible(False)
plt.subplots_adjust(left = 0.02, bottom = 0.02, right = 0.98, top = 0.98, wspace = 0.10)
ax[0].set_title("DigitalGlobe")
ax[1].set_title("OpenStreetMap")
fig.suptitle("Building Count Choropleth for Freetown, SL")
# plt.savefig("n_choropleth_sl.png", dpi=300, bbox_inches="tight")
plt.show()
# get block-by-block diff of k values
diff = dg.join(osm, rsuffix="_osm")
diff["delta_n"] = diff["num_buildings"] - diff["num_buildings_osm"]
diff.plot("delta_n", legend=True)#, cmap="PRGn", scheme = "natural_breaks", missing_kwds={"color": "lightgrey"})
plt.gca().get_xaxis().set_visible(False)
plt.gca().axes.get_yaxis().set_visible(False)
plt.title(r'Block-by-block Difference in Building Count ($n_{DG} - n_{OSM}$)')
plt.subplots_adjust(left = 0.02, bottom = 0.02, right = 0.98, top = 0.88, wspace = 0.10)
# plt.savefig("delta_n_choropleth_sl.png", dpi=300, bbox_inches="tight")
plt.show()
def visualize_complexity(dg: gpd.GeoDataFrame, osm: gpd.GeoDataFrame):
# get k range
max_k = max(dg.complexity.max(), osm.complexity.max())
min_k = 0
# visualize complexity distribution
fig, ax = plt.subplots(1, 2, sharey='row', sharex='col')
bins = range(0, max_k + 1, 1)
dg.complexity.hist( ax = ax[0], bins = bins)
osm.complexity.hist(ax = ax[1], bins = bins)
ax[0].set_title("DigitalGlobe")
ax[1].set_title("OpenStreetMap")
fig.suptitle("Histogram of $k$-values for Freetown, SL")
plt.semilogy()
# plt.savefig("k_histogram_sl.png", dpi=300, bbox_inches="tight")
plt.show()
# complexity choropleth
fig, ax = plt.subplots(1, 2, sharey='row', sharex='col')
plot_opts = {
"column" : "complexity",
"cmap" : mnp_palette,
# "scheme" : "natural_breaks",
"scheme" : "User_Defined",
"edgecolor": "grey",
"linewidth": 0.1,
"legend" : True,
"categorical" : False,
"markersize" : 0,
"classification_kwds" : {"bins" : [6, 15, 31, 76, max_k]} # numbers taken from natural_breaks
}
dg.plot( ax=ax[0], **plot_opts)
ax[0].get_xaxis().set_visible(False)
ax[0].axes.get_yaxis().set_visible(False)
# dummy polygons to make choropleth color scheme consistent
osmd = gpd.GeoDataFrame(pd.concat([pd.DataFrame([
["d0", shapely.geometry.MultiPoint(osm.centroid).centroid, 0, [], 0],
["d1", shapely.geometry.MultiPoint(osm.centroid).centroid, 2, [], 0],
["d2", shapely.geometry.MultiPoint(osm.centroid).centroid, 5, [], 0],
["d3", shapely.geometry.MultiPoint(osm.centroid).centroid, 19, [], 0],
["d4", shapely.geometry.MultiPoint(osm.centroid).centroid, max_k, [], 0]
], columns = osm.columns), osm.copy()]).reset_index(), geometry = "block_geom")
osmd.plot(ax=ax[1], **plot_opts)
ax[1].get_xaxis().set_visible(False)
ax[1].axes.get_yaxis().set_visible(False)
plt.subplots_adjust(left = 0.02, bottom = 0.02, right = 0.98, top = 0.98, wspace = 0.10)
ax[0].set_title("DigitalGlobe")
ax[1].set_title("OpenStreetMap")
fig.suptitle("Complexity Choropleth for Freetown, SL")
# plt.savefig("k_choropleth_sl.png", dpi=300, bbox_inches="tight")
plt.show()
# get block-by-block diff of k values
diff = dg.join(osm, rsuffix="_osm")
diff["delta_k"] = diff["complexity"] - diff["complexity_osm"]
diff.plot("delta_k", legend=True, cmap="PRGn", scheme = "natural_breaks", missing_kwds={"color": "lightgrey"})
plt.gca().get_xaxis().set_visible(False)
plt.gca().axes.get_yaxis().set_visible(False)
plt.title(r'Block-by-block Difference in Complexity Measure ($k_{DG} - k_{OSM}$)')
plt.subplots_adjust(left = 0.02, bottom = 0.02, right = 0.98, top = 0.88, wspace = 0.10)
# plt.savefig("delta_k_choropleth_sl.png", dpi=300, bbox_inches="tight")
plt.show()
