Property data that is part document, part network
Skythena aggregates real estate data from multiple sources and serves analytics through REST APIs. Property attributes vary widely, which called for flexible document storage, while ownership chains, geographic proximity and market comparisons form a connected network best modelled as a graph.
Two databases, each for what it does best
We built the platform on MongoDB for property documents and Neo4J for the relationships between properties, owners and market segments. It also processes geospatial real estate data.
The API layer, MongoDB access and Neo4J graph queries each run as a separate service on their own AWS EC2 instances, so each kind of query gets a data tier built for it and the infrastructure scales.
What it makes possible
With document storage for flexible property schemas and graph queries for relationship analysis, Skythena runs analytics that would be hard to achieve on a single database technology.