Jachym
Czech GIS ex-celebrity that is constantly facing burnout. Ex-forestry expert who hit GIS and programming by accident and never fully recovered. Started to be high on GRASS, moved to the backend as the lead of the PyWPS project, and on the frontend side had to fight with various JavaScript libraries (you name it) in order to build spatial data portals. Currently retreating as (mostly raster) data analyst, where writing a few lines of code usually fully utilize a full stack of servers for a couple of days - so you can re-run the script after it’s done with correct input parameters while the result is false: repeat. OSGeo and FOSS4G community member. GeoData team at MapTiler!
Sessions
When publishing (raster and vector) data in the form of a web mapping application, the first step is always to prepare a cache of the data. Currently, tiled images seem to be the industry standard - and the internal format of the tiles is either PBF (for vector data) or PNG/JPEG/WebP or similar raster data formats supported by current web browsers and desktop mapping applications (e.g. QGIS).
Most of the tools out there are going to store the raster tiles in a file-system structure, using directories for the Z and X tile coordinates and file names for the Y coordinate. This is limiting for practical purposes as on some filesystems you can exceed the maximum number of files easily. While for the vector data, the OpenMapTiles project seems to be well established, along with Tippecanoe and Planetiler, for the raster data tiles, the field of tiling possibilities is wide open.
The tiling process can be very demanding on hardware resources and time-consuming. Having the possibility to parallel process the data or even use a cluster of machines for faster tiling could be crucial for some applications.
In this talk, we will give an overview of the current possibilities for tiling, focused (but not exclusively) on the raster data tiles. Gdal2tiles, QGIS tile generating tools, mapproxy-seed, mapcache_seed, and others. Each of the tools has its place in the geospatial data provider ecosystem, and so does MapTiler-Engine. With MapTiler-Engine, users can process large amounts of geospatial data and store them in various output tile formats. It supports many input data formats and adds modifications such as output color, resolution, and more. It also supports different tile matrix sets. MapTiler-Engine has a graphical user interface for easy usage, but it also has a command line interface, so you can make it part of a larger toolchain.