Skip to content

Latest commit

 

History

History
130 lines (77 loc) · 3.78 KB

STEPS.md

File metadata and controls

130 lines (77 loc) · 3.78 KB
Raw

Import regions from OpenStreetMap

The total process takes about 1 hour.

Prerequisites

Land polygons

Download "WGS 84, Large polygons not split" from osmdata, unpack and store the shapefile in

  • 00-static-data/land-polygons-complete-4326

Direct link: https://osmdata.openstreetmap.de/download/land-polygons-complete-4326.zip

Landing page: https://osmdata.openstreetmap.de/data/land-polygons.html

You may want to open the file in mapshaper and check that the polygons do not self-intersect, because clipping with self-intersecting polygons will corrupt data. For example, clipping with these polygons removes most of Japan from the output:

Screenshot

You can also use this snapshot of land polygons with correct polygons.

Natural earth

Natural earth data (1:10m Cultural Vectors) is used for zoom levels 0 to 4 for compatibility with OpenMapTiles.

Download "countries" and "states and provinces" and unpack and store the shapefiles in

  • 00-static-data/ne_10m_admin_0_countries
  • 00-static-data/ne_10m_admin_1_states_provinces

Direct links:

Landing page: https://www.naturalearthdata.com/downloads/10m-cultural-vectors/

Dependencies

Run

npm install

to install the dependencies.

Steps

Run this script to execute all steps listed below:

import-osm/import-osm.sh

1. Query list of countries (Overpass)

Input: Nothing. Output: List of countries with ISO3166-1 codes.

2. Query regions by country (Overpass)

Input: List of countries with ISO3166-1 codes. Output: For every country, one GeoJSON file with country and subdivision polygons.

Also store raw data only download if raw data is not available.

3. Clip with land polygons

Input: For every country, one GeoJSON file with country and subdivision polygons. Output: For every country, one GeoJSON file with country and subdivision polygons.

4. Reduce regions (remove small disconnected parts, e.g. remove French Guiana from France)

Input: For every country, one GeoJSON file with country and subdivision polygons. Output: For every country, one GeoJSON file with country and subdivision polygons.

5. Split by region

Input: For every country, one GeoJSON file with country and subdivision polygons. Output: For every region, one GeoJSON file.

6. Simplify regions

Input: For every region, one GeoJSON file. Output: For every region, one GeoJSON file.

7. Merge regions

Input: For every country, one GeoJSON file with country and subdivision polygons. Output: One GeoJSON file with all countries, one GeoJSON file with all subdivisions.

8. Generate vector tiles

Input: One GeoJSON file with all countries, one GeoJSON file with all subdivisions. Output: mbtiles file with 2 layers (countries, subdivisions).

9. Convert natural earth data to GeoJSON

Input: Shapefiles with countries and states/provinces. Output: GeoJSON files with countries and states/provinces.

10. Generate vector tiles (natural earth)

Input: GeoJSON files with countries and states/provinces. Output: mbtiles file with 2 layers (countries, subdivisions).

11. Join tiles

Input: mbtiles files from steps 8/10. Output: mbtiles file with 2 layers (countries, subdivisions), using natural earth data for zoom levels 0-4 and Openstreetmap data for zoom levels 5-10.

Clean up

Run this to remove all output folders:

import-osm/remove-outputs.sh

Preview vector tiles

Run this script to preview the vector tiles generated in step 11.

import-osm/preview-tiles.sh