22 April 2007

Freebase !

I finaly received my invitation to join freebase. Freebase ,which was previously introduced by Tim O'Reilly, is a structured semantic wiki. I tested it today: this is a great product. I consider the whole site as a RDF/RDFS editor where you can define classes , properties and create some instances. In consequence, Freebase is far more structured than wikipedia.

The site comes with a complete API (MQL Metaweb Query Language(looks like SPARQL)) which can be used to query freebase and to create your own application (e.g. see CineSpin).

Example: searching for physicists born between 1800 and 1900:


{
"query":[{
"/people/person/date_of_birth":null,
"/people/person/date_of_birth<":"1900",
"/people/person/date_of_birth>=":"1800",
"limit":35,
"name":null,
"type":"/science/physicist"
}]
}


Result:


{
"result":[{
"/people/person/date_of_birth":"1879-03-14",
"name":"Albert Einstein",
"type":"/science/physicist"
},{
"/people/person/date_of_birth":"1878",
"name":"Lise Meitner",
"type":"/science/physicist"
},{
"/people/person/date_of_birth":"1867",
"name":"Marie Curie",
"type":"/science/physicist"
},{
"/people/person/date_of_birth":"1844",
"name":"Ludwig Boltzmann",
"type":"/science/physicist"
}],
"status":"/mql/status/ok"
}


Considering bioinformatics many types could be created (I've no time to play with this at this time ! ): defining Molecular Interactions, Biologists, etc...

Pierre

2 comments:

Anonymous said...

Pierre

Glad to hear you got an invite. It will be fun to see what you can do with it. I can't wait for my workload to lighten up, so that I can do more during the beta test phase

Pedro Beltrao said...

I have an invite too. I only had a quick look at it. I am unfortunately to busy right now to have a go at it but it looks great. I had the feeling of a kid in the toy store. There is so much to do in there, so much territory to map. They should get people like Nicolas Le Novere and Kitano involved it setting up the standards.
There is just the little problem of the copyright on the data. Somethings I don't think we can upload there (SNPs, protein interactions, etc).
The potential for collaboration is huge. Most of the boring job in bioinformatics is getting hold of the data, crossing flat files, parsing XML,etc. If a lot of people used freebase to store their data (specially the important databases) itwould cut down the time and boredom of getting the data.