FOSS4G 2022 general tracks

Nick Peihl


Sessions

08-25
10:05
5min
What’s new in geospatial Elasticsearch
Jorge Sanz, Nick Peihl, Craig Taverner

Elasticsearch is a well-known and mature NoSQL database providing search and analytics services for big datasets. The “elasticity” of its name comes from the distributed design and easy scalability capabilities that have made it an industry leader for more than ten years. In this talk we will present two exciting new features that have been added recently to the product related with the geospatial topic: vector tiles support and line and hexagon aggregations.

Vector tiles have become an industry standard to encode large amounts of data to be displayed in the browser by web mapping libraries like MapLibre or OpenLayers. Elasticsearch analytics & geo team has added a new API endpoint that renders search and aggregation queries as zipped protobuffers, allowing developers to retrieve right from the datastore assets that are ready to be sent to the user's browser without much further processing. This will speed up the rendering of large datasets by avoiding transferring JSON assets from Elasticsearch to application middleware.

Elasticsearch geospatial aggregation capabilities have been extended recently by two new methods, one is to allow combining related points into a new line geometry (think of a vehicle track) and the other is to aggregate geometries into an hexagon grid. The new geo-line aggregation will be very useful for asset tracking use cases where the second enables Elasticsearch to perform powerful analytics combined with the extensive support for metric aggregations.

In this talk we will present this project, going through the different use cases with some examples and demonstrations using both Kibana Elastic Maps and a simple ad-hoc web project that leverages this new feature.

State of software
Room 4