Skip to content

Talk · Knowledge Graph Conference 2023

Graph.Build.Done

Richard Loveday, Head of Product, at the Knowledge Graph Conference in New York. Building a knowledge graph end to end on stage, from source data to a loaded database, in the time a conference slot allows.

The recording

From a source system to a working graph, live

The argument in the talk is the one the platform is still built on: the expensive part of a knowledge graph is not the database, it is everything between the source data and the model, and that part can be automated.

  • Knowledge Graph Conference, New York, 2023
  • Richard Loveday, Head of Product
  • Studio has moved on since. The argument has not.
graph.build.done · KGC23 20 min · new york · 2023
Recorded at the Knowledge Graph Conference 2023.

Since then

Three years of it, in the places it now lives

The talk is from 2023. If it made the case, these are the current versions of the three things it demonstrated.

Studio

The modelling environment from the talk, now git-backed, able to import external ontologies and to model RDF and property graphs from a single description.

Transformers

The ETL. More sources than the talk had, including Kafka, REST and documents, and mappings that are open RML and R2RML rather than a private format.

See it on your own data

A recording of somebody else's domain only goes so far. Forty-five minutes on yours, with nothing installed, settles more than an hour of video.