01_convert.py

import h3
import json
import os
import pandas as pd
import geopandas as gpd
from kml2geojson import main as k2g


# Convert Boundary Data
zones = k2g.convert('boundaries/bbmp_zone_boundaries_243.kml')
with open('boundaries/zones.geojson', 'w') as f:
    f.write(json.dumps(zones[0]))

wards = k2g.convert('boundaries/bbmp-final-new-wards.kml')
with open('boundaries/wards.geojson', 'w') as f:
    f.write(json.dumps(wards[0]))
 

# Convert Census Data
# ___

def point_to_h3(obj):
    idx = str(h3.latlng_to_cell(obj.y, obj.x, 8))
    return idx

RAW_DATA_DIR = 'raw_census_data/'
gdfs = []

for filename in os.listdir(RAW_DATA_DIR):
    print(f'READING {filename}')
    ds = k2g.convert(RAW_DATA_DIR + filename)
    zone_gdf = gpd.GeoDataFrame.from_features(ds[0]['features'])
    zone_code = filename.split('blr_')[1].split("_zone")[0].lower()
    zone_gdf['Zone'] = zone_code
    zone_gdf['H3Index8'] = zone_gdf['geometry'].apply(point_to_h3)
    gdfs.append(zone_gdf)
    

gdf = pd.concat(gdfs)
gdf.to_file('census_data_converted.geojson', driver="GeoJSON")