Wednesday, September 1, 2010

Freebase at Google

Kirrily Robert of Freebase (and now Google) gave a presentation at this month's GTUG meeting, describing the technology stack, queries, and data management.

This is a collection of information gathered at the talk, though it is mostly just an amalgamated Twitter feed. Still, there are links to all the resources one would need to get starting programming against the Freebase platform.

Freebase has been around for years and was recently acquired by Google. It uses a semantic web model of linked data with dynamic ontologies, more or less. It features a public REST API whose parameter is a structured query (which can be treated as a subset of SPARQL) and returns results in JSON format. It can also return results in RDF.
Twitter: @fbase

one example of a site using Freebase data is

Metaweb Query Language MQL (pronounced like pickle) used to access Freebase data

Freebase MQL Query Editor available online

best supported library is freebase-python and there are others too

Online Freebase App Editor

Acre is also a server-side javascript application framework for Freebase

Freebase 102 demo application done using Acre

Full data dumps of every fact and assertion in Freebase are available weekly

the Freebase RDF Service

RABJ (pronounced like cabbage) Redundant Array of Brains in a Jar

Freebase Gridworks for dealing with messy tabular data and cleaning it up

Kirrily Robert @skud runs a regular Freebase meetup in San Francisco too

OpenCalais can give you freebase identifiers as part of its analysis. Some news organizations (NYTimes, UK Guardian) have built linked data APIs which can be integrated in a mash-up with Freebase or other linked data applications.