Docker image containing geospatial data for Protx portal.
How it is used | Additional notes | ||
---|---|---|---|
geojson files | served via nginx to frontend client | ||
vector tiles | served via nginx to frontend client | These are uncompressed vector tilest | |
PostgreSQL/PostGIS database | utilized by protx backend to create some responses (e.g. resource downloads) |
This image is Used by the following repos:
To build and deploy to dockerhub:
make build
make publish
The initial work is described in confluence.
The basic steps are:
- convert shapefiles to geojson (using gdal's
ogr2ogr
) - use
tippecanoe
to create static vector tiles
These tippecanoe uncompressed files work but there are some shortcomings: (i) frontend gets a lot of missing errors when zooming to places outside the bounds of our generated vector tiles and (ii) the size of the data is large. COOKS-27 details a possible approach to improve things by serving the tiles (see ticket for more info: "To create mbtiles and then serve them through tileserver-gl (TileServer-GL, https://github.com/klokantech/tileserver-gl)) . Or there are other ways to do this (postgis?)
We need a geospatial database (postgis) and endpoint to handle queries like (1) searching through names (e.g. is any county named "Bexar"?) and (2) what is the extent of my selected region (e.g. what is the outline of this county)
This folder contains the original data used in this docker image:
texas_counties.shp
is from/corral-secure/projects/Cooks-ProTX/spatial/tigris/texas_counties/
Texas_State_Boundary.geojson
is from https://gis-txdot.opendata.arcgis.com/datasets/texas-state-boundary/explore
Note: In early versions of the portal, we were also presenting census tracs, urban areas, zip codes and dfps regions. Currently, we are only using county data with an outline of texas but a complete list of other geographic regions used in earlier demos can be found in confluence.