View Source

{toc}

h1. Freebase

Freebase is a crowdsourced knowledge base with very flexible schema and simple structuring. Bought by Google.
- I now read several presentations and howto's.
- I'm sure we can take away a lot of useful editing examples.
- Powerful query editor and language (MQL):
[http://www.freebase.com/queryeditor/about]
- 2009 presentations with very impressive screenshots
[http://www.slideshare.net/metawebrobert/ny-semantic-web-meetup-27-aug-2009]
[http://www.slideshare.net/metawebrobert/ny-freebase-workshop-10-dec-2009]
-- demoes wide coverage: screen-fulls of actors, cat breeds, beers, mountains...
-- semantic searches by Microsoft Bing ([paintings!|http://www.bing.com/images/search?q=rembrandt+paintings&qpvt=rembrandt+paintings&FORM=IQFRML]) *and* Google Labs
-- collection search (eg "architects of sckyscrapers in Hong Kong")
-- suggest/autocomplete showing nice description of each entity
- ACRE: web-based app dev platform
- link graphs (eg Popstra), etc etc

ACRE server-side JS environment on Freebase:
- [http://www.slideshare.net/WillMoffat/serverside-javascript-on-freebase-sf-javascript-meetup-9]

I think that we definitely must include FreeBase and relatives in our RS4 explorations.

h2. Freebase data
We have experience with Freebase data (though not with the application):
- Our [http://www.FactForge.net] (FF) incorporates FreeBase data
- FB doesn’t use RDF constructs (e.g. rdf:type) nor RDFS/OWL inference
- We're partners with Google in the RENDER project
- As part of this, we're converting FB's constructs to standard RDFS/OWL constructs. FF includes a mapping layer (through Proton) that assigns standard rdf:type to FB data
- We used FB food data for mapping recipe ingredients to the USDA Standard Reference database in the [Edamam project|http://www.ontotext.com/case/pagane-foodKB]


h1. Parallax

[http://www.freebase.com/labs/parallax/]
Parallax is an alternative search interface for FB. It is based on a "set-based browsing" paradigm, and exploring
- Please watch the video (Aug 2008). Very inspiring, and made me jealous :-)
[http://vimeo.com/1513562]
- That’s made 4y ago and is open sourced. I imagine a lot more's been done since Google purchased them.
- [Parallax and Companion- Set-based Browsing for the Data Web (WWW 2009)|http://davidhuynh.net/media/papers/2009/www2009-parallax.pdf] by David Huynh
- Note: I tried making a 2D chart of 3700 skyscrapers (Height vs Construction Cost) but it locked the browser.
- Wanted to try on fewer (highest 100), but when I added "More filters> Height", the Settings link didn't do anything.

h2. Companion
TODO: need to explore