FOSS4G 2022 general tracks

TiTiler: Not just a tile server
2022-08-25, 09:00–09:30 (Europe/Rome), Modulo 0

During the last 2 years we've been working on TiTiler (https://developmentseed.org/titiler/), a dynamic raster tile server. Built on top GDAL/Rasterio, TiTiler is written in python and use FastAPI (https://fastapi.tiangolo.com) framework. TiTiler is an application that let you create raster tiles dynamically from raster datasets (e.g Cloud Optimized GeoTIFF) but also from Spatial Temporal Asset Catalog (STAC) or Mosaic (using MosaicJSON). It is also a set of python modules which can be used independently to create custom services.

During this talk we'll explain the concept of dynamic tiling, what is TiTiler (and the libraries powering it), how it works and more important how users can customize and built their own dynamic tile server.

We will also present project like TiTiler-PgSTAC (https://github.com/stac-utils/titiler-pgstac) which enables the creation of Mosaic tiles dynamically from a Spatial Temporal Asset Catalog (STAC) database, or eoAPI (https://github.com/developmentseed/eoAPI) which is a full Earth Observation data service combining STAC database, STAC-FastAPI and a TiTiler in one easily deployable project.

Vincent is a geospatial developer at Development Seed. He builds open source tools, infrastructure and web interfaces that process and visualize geospatial data. The useable satellite interfaces he builds help break down the barrier to accessing and exploring Earth observation data.

Vincent is also heavily involved in the Cloud Optimized GeoTIFF (a.k.a COG) specification and communication. He is the main contributor on multiple open-source project to create / use / analyse COG such as rio-cogeo, rio-tiler or tilebench. He is also involved in other open-source project such as Rasterio or STAC-FastAPI.

Prior to joining Development Seed, Vincent created the website RemotePixel.ca to develop web interface and cloud processes to access satellite data such as Landsat and Sentinel and distributes the data through simple and intuitive interfaces. Vincent previously worked as a geospatial developer at Mapbox.