ExternalVendorCode/tippecanoe
1
0
Fork 0
mirror of https://github.com/mapbox/tippecanoe.git synced 2025-01-22 04:18:01 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
Build vector tilesets from large collections of GeoJSON features.
159 Commits 74 Branches 87 Tags 45 MiB
C++ 95%
Makefile 2.5%
C 1.7%
Perl 0.5%
Shell 0.3%
69fa9e5f22
Go to file
Eric Fischer 69fa9e5f22 Remove jsoncat since it doesn't belong in this project
2014-09-26 16:14:13 -07:00
geojson.c Restrict to one input file, since that's all that will work 2014-09-26 14:33:03 -07:00
jsonpull.c Preserve the original string representation of numbers too. 2014-02-26 17:01:30 -08:00
jsonpull.h Clean up after errors and fix related bugs: 2014-02-08 09:31:49 -08:00
LICENSE.md Add license 2014-02-08 10:33:57 -08:00
Makefile Fix build errors on Mac 2014-09-26 15:25:12 -07:00
README.md More README clarification 2014-09-26 15:36:41 -07:00
tile.cc Remember to finalize statements 2014-09-26 14:51:29 -07:00
tile.h Actually write tiles directly to the .mbtiles file! 2014-09-25 16:34:17 -07:00
vector_tile.proto Add vector tile boilerplate 2014-09-22 10:45:34 -07:00

README.md

tippecanoe

Build vector tilesets from large collections of GeoJSON features.

Usage

tippecanoe -o file.mbtiles [file.json]

If the file is not specified, it reads GeoJSON from the standard input.

Options

  • -l Layer name (default "file" if source is file.json)
  • -n Human-readable name (default file.json)
  • -z Base zoom level (default 14)
  • -Z Lowest zoom level (default 0)
  • -d Detail at base zoom level (default 12, for tile resolution of 4096)
  • -D Detail at lower zoom levels (default 10, for tile resolution of 1024)

Example

tippecanoe -o alameda.mbtiles -l alameda -n "Alameda County from TIGER" -z12 -d14 tl_2014_06001_roads.json

Geometric simplifications

At every zoom level, line and polygon features are subjected to Douglas-Peucker simplification to the resolution of the tile.

For point features, it drops 1/2.5 of the dots for each zoom level above the base. I don't know why 2.5 is the appropriate number, but the densities of many different data sets fall off at about this same rate.

For line features, it drops any features that are too small to draw at all. This still leaves the lower zooms too dark (and too dense for the 500K tile limit, in some places), so I need to figure out an equitable way to throw features away.

It also throws away any polygons that are too small to draw. I'm not sure yet if it is appropriate to do more than that.

Development

Requires protoc (brew install protobuf), protobuf-lite, and sqlite3. To build:

make

and perhaps

make install