WS35: Building geo search applications with Elasticsearch

Trainers: Mark Varley (AddressCloud)

Schedule: Tuesday Noon 14-18

Places: 30

Abstract:

In this workshop we will explore the background of NoSQL databases, why they exist and which use cases they are particularly suitable for. We will look at some key considerations such as eventual consistency to understand the trade offs required when designing applications with this technology. We will then explore various different approaches and open source solutions available and the relative merits of each. For the remainder of the workshop we will focus on Elasticsearch, an open source distributed, NoSQL database and full-text search engine and in particular its geospatial capabilities which have recently had a major upgrade. We will follow a case study with worked examples taking a large dataset, indexing it in Elasticsearch, looking at the various options for querying including filters and full text queries and completion suggesters and then drill in on the geospatial queries, filters and sorting that are available. We will conclude by discussing deployment strategies and techniques for the real world to avoid nasty surprises in production.

Tags: elasticsearch, spatial index, NoSQL

Additional Information:

Elasticsearch is being adopted by organisations from NASA to the USGS to investment banks to small startups such as AddressCloud. However, some see this is a silver bullet and the workshop leader will help attendees understand when it makes sense to use this and when, for example, it would make more sense to use a more appropriate solution such as full text search in PostGIS. As well as the theory attendees will take part in worked examples and come away with a good understanding and some real tangible skills.