GeoJSON -> MBTiles with Supercluster
npm install supertilerUse Supertiler when you want to make a tiled dataset out of a large set of GeoJSON point features. Supertiler uses Supercluster to cluster points, in order to improve performance and keep low-zoom tiles from violating the 500KB size limit for tilesets hosted by Mapbox. For a more general purpose tileset generating tool, see Tippecanoe. Among its many features, Tippecanoe also supports point clustering, but using Supercluster has a few advantages:
- You can control the order in which features are added to clusters. This is helpful for making clusters tend to be geographically centered around the "most important" of their features.
- Supercluster's cluster radius is precise, while Tippecanoe uses a (faster) approximation
- Supercluster supports custom property aggregation functions, while Tippecanoe provides a set of pre-defined aggregators.
- Mapbox GL JS uses Supercluster for clustering, so if you've made a map with client-side clustering and want to switch to server-hosted (and tiled) data, Supertiler allows for a direct translation.
Many thanks to @mourner and @ericfischer, from whom I have copied liberally!
import supertiler from 'supertiler';
supertiler({
input: 'input.geojson',
output: 'output.mbtiles',
maxZoom: 2
}).then(null, err => console.log(err));
$3
npm install -g supertiler
supertiler -i input.geojson -o output.mbtiles
Usage: supertiler.js [options]Options:
--version Show version number [boolean]
--input, -i Input in GeoJSON format. Each feature's geometry
must be a GeoJSON Point. [required]
--output, -o Path of output MBTiles database [required]
-h, --help Show help [boolean]
--includeUnclustered Include one zoom level above the max clustering
zoom. [boolean] [default: true]
--minZoom Minimum zoom level at which clusters are
generated. [number] [default: 0]
--maxZoom Maximum zoom level at which clusters are
generated. [number] [default: 8]
--radius Cluster radius, in pixels. [number] [default: 40]
--extent (Tiles) Tile extent. Radius is calculated
relative to this value. [number] [default: 512]
--nodeSize Size of the KD-tree leaf node. Affects
performance. [number] [default: 64]
--map A javscript function that returns cluster
properties corresponding to a single point. See
supercluster documentation. [string]
--reduce A javscript reduce function that merges
properties of two clusters into one. See
supercluster documentation. [string]
--filter A javscript function that filters features from
the resulting tiles based on their properties.
[string]
--storeClusterExpansionZoom Store the "cluster expansion zoom" as a property
for each cluster. [boolean] [default: false]
--attribution (HTML): An attribution string, which explains the
sources of data and/or style for the map.[string]
--bounds (string of comma-separated numbers): The maximum
extent of the rendered map area. See MBTiles
spec. [string] [default: (entire world)]
--center (comma-separated numbers): The longitude,
latitude, and zoom level of the default view of
the map. See MBTiles spec.
[string] [default: (null island)]
--description A description of the tileset's content. [string]
--tileSpecVersion The version of the Mapbox Vector Tile
Specification to use. See MBTiles spec.
[number] [default: 2]
--logPerformance Output performance timing logs to console.
[boolean]
Examples:
supertiler.js -i in.geojson -o Cluster from zoom 0 to 5 while
out.mbtiles --minZoom 0 --maxZoom 5 aggregating "myValue" into "sum"
--map "(props) => ({sum: property. Outputs tileset with zoom
props.myValue})" --reduce "(accumulated, from 0 to 6.
props) => { accumulated.sum +=
props.sum; }"
Generation will fail if any tile goes over maximum size of 500KB. In this case,
try increasing cluster radius, increasing max zoom, or generating fewer
aggregated properties.
Developing
npm install # install dependencies
npm test # run tests