The maps here were created using the following steps.
Census files have to be downloaded from a browser. This assumes they are downloaded to ~/Downloads. us-state/zip/county Available: http://www.census.gov/cgi-bin/geo/shapefiles2010/main
- Open a browser to: https://www.census.gov/cgi-bin/geo/shapefiles2010/main
- Choose 'States (and equivalent)' and 'Submit'
- Choose 'All states in one national file' for 'State and Equivalent (2010)' and click 'Download'
Then unzip the files:
mkdir -p raw/state && cd raw/state && unzip ~/Downloads/tl_2010_us_state10.zip && cd ../..
- Open a browser to: https://www.census.gov/cgi-bin/geo/shapefiles2010/main
- Choose 'Counties (and equivalent)' and 'Submit'
- Choose 'All states in one national file' for 'County and Equivalent (2010)' and click 'Download'
Then unzip the files:
mkdir -p raw/county && cd raw/county && unzip ~/Downloads/tl_2010_us_county10.zip && cd ../..
- Open a browser to: https://www.census.gov/cgi-bin/geo/shapefiles2010/main
- Choose 'Zip Code Tabulation Areas' and 'Submit'
- Choose 'All states in one national file' for '5-Digit ZIP Code Tabulation Area (2010)' and click 'Download'
Then unzip the files:
mkdir -p raw/zcta5 && cd raw/zcta5 && unzip ~/Downloads/tl_2010_us_zcta510.zip && cd ../..
Medicare files can be downloaded directly from Dartmouth Atlas hrr/hsa Available: http://www.dartmouthatlas.org/Tools/Downloads.aspx?tab=35
curl -O http://www.dartmouthatlas.org/downloads/geography/hrr_bdry.zip
mkdir -p raw/hrr && cd raw/hrr && unzip ../../hrr_bdry.zip && cd ../..
curl -O http://www.dartmouthatlas.org/downloads/geography/hsa_bdry.zip
mkdir -p raw/hsa && cd raw/hsa && unzip ../../hsa_bdry.zip && cd ../..
Clean up .zip files as needed.
Convert using gdal. On a Mac, you can install gdal with homebrew: brew install gdal.
ogr2ogr -f "GeoJSON" hsa.json ./raw/hsa/HSA_Bdry.SHP HSA_Bdry
ogr2ogr -f "GeoJSON" hrr.json ./raw/hrr/HRR_Bdry.SHP HRR_Bdry
ogr2ogr -f "GeoJSON" state.json ./raw/state/tl_2010_us_state10.shp tl_2010_us_state10
ogr2ogr -f "GeoJSON" county.json ./raw/county/tl_2010_us_county10.shp tl_2010_us_county10
ogr2ogr -f "GeoJSON" zcta5.json ./raw/zcta5/tl_2010_us_zcta510.shp tl_2010_us_zcta510
Move the GeoJSON files into the geojson directory
mkdir -p geojson && mv *.json geojson
TopoJSON is an extension to GeoJSON that encodes toplogy, resulting in much smaller files. Nodejs is required to run the script. Install topojson using npm install topojson -g. No simplification is done on these files. None of the files are combined either, although that is possible. ZCTA file fails when trying to use topojson to convert.
mkdir -p topojson
topojson --properties HRRCITY --id-property HRRNUM --out ./topojson/hrr.json ./geojson/hrr.json
topojson --properties HSANAME --id-property HSA93 --out ./topojson/hsa.json ./geojson/hsa.json
topojson --properties NAME10 --properties STUSPS10 --id-property GEOID10 --out ./topojson/state.json ./geojson/state.json
topojson --properties NAME10 --properties NAMELSAD10 --id-property GEOID10 --out ./topojson/county.json ./geojson/county.json
topojson --id-property GEOID10 --out ./topojson/zcta5.json ./geojson/zcta5.json
I wanted some of these files combined into a single TopoJSON, and simplified somewhat to make the size of the map smaller. To do so, the following command was used:
topojson --out ./topojson-simplified/us.json ./topojson/state.json ./topojson/hrr.json ./topojson/county.json
topojson -s 1 --out ./topojson-simplified/us-s1.json ./topojson/state.json ./topojson/hrr.json ./topojson/county.json
topojson -s 3 --out ./topojson-simplified/us-s3.json ./topojson/state.json ./topojson/hrr.json ./topojson/county.json
A simplification threshold of 3 retained 150581 / 549117 points (27%), while a level of 1 retained 231768 / 549117 points (42%).
The reference directory has lookup files to get names for each of the IDs, as well as getting parent information (e.g. which state is a HRR in).