Feature request: add parcel development risk

[nlebovits]

Some users are interested in the likelihood that a vacant property will be developed in the near future. We could build a model to predict this, but the number of building permits issued in the past year in the Census block group in which the parcel sits is a pretty good proxy. To that end, we should add these data in an interpretable way.

Describe the solution you'd like
We should add a count of the building permits issued per block group in the past year; classify it into high, medium, and low using Jenks breaks; and spatially join this to properties.

Describe alternatives you've considered
Alternatively, we could build some kind of predictive model for this, but that's outside the immediate scope of this project and more effort than it's worth for our purposes.

Additional context
Code to implement this is here:

import pandas as pd
import geopandas as gpd
import matplotlib.pyplot as plt
import requests
from datetime import datetime, timedelta
import jenkspy



census_bgs_url = 'https://opendata.arcgis.com/datasets/2f982bada233478ea0100528227febce_0.geojson'
census_bgs_gdf = gpd.read_file(census_bgs_url)


base_url = "https://phl.carto.com/api/v2/sql"

one_year_ago = (datetime.now() - timedelta(days=365)).strftime('%Y-%m-%d')

query = f"""
         SELECT
         permitissuedate,
         the_geom,
         ST_AsGeoJSON(the_geom)::json AS the_geom_geojson
         FROM permits
         WHERE permitissuedate >= '{one_year_ago}'
         """

response = requests.get(f"{base_url}?q={query}&format=GeoJSON")

permits_gdf = gpd.GeoDataFrame.from_features(response.json(), crs='EPSG:4326')

if census_bgs_gdf.crs != permits_gdf.crs:
    permits_gdf = permits_gdf.to_crs(census_bgs_gdf.crs)

joined_gdf = gpd.sjoin(permits_gdf, census_bgs_gdf, how="inner", op='within')

permit_counts = joined_gdf.groupby('index_right').size()

census_bgs_gdf['permit_count'] = census_bgs_gdf.index.map(permit_counts)
census_bgs_gdf['permit_count'] = census_bgs_gdf['permit_count'].fillna(0)

breaks = jenkspy.jenks_breaks(census_bgs_gdf['permit_count'], n_classes=3)
census_bgs_gdf['dev_rank'] = pd.cut(census_bgs_gdf['permit_count'],
                        bins=breaks,
                        labels=['low', 'medium', 'high'])