Freebasing
June 12, 2013 11:41 PM   Subscribe

Freebase, formerly Metaweb, is a Google-run open, community-curated database of everything: 40 million topics and over a billion facts, all free to use. If you need to wrangle the resulting data into another format, Mr Data Converter (previously) has you covered.
posted by Bora Horza Gobuchul (23 comments total) 29 users marked this as a favorite
 
I'm glad someone is still on board with the idea of a semantic web, and for all their flaws I'm unsurprised it's Google.
posted by jaduncan at 11:47 PM on June 12, 2013 [3 favorites]


Google has announced as of January 1st, 2014, this service will no longer be available.
posted by [insert clever name here] at 1:47 AM on June 13, 2013 [9 favorites]


Man, this site is so addictive, for some reason...
posted by ShawnStruck at 2:27 AM on June 13, 2013 [1 favorite]


Google has announced as of January 1st, 2014, this service will no longer be available.

Citation? I'm assuming this is a joke given the shoutouts at IO, but wanted to check.
posted by jaduncan at 2:31 AM on June 13, 2013


Music- 25 MILLION Facts.
Physics- 55 facts.

This is what happens when the internets takes over.


Given, the music stuff looks like it could all be cribbed automatically from Discogs.
posted by solarion at 2:33 AM on June 13, 2013


Interestingly enough, Freebase is the corpus that the image recognition stuff in G+ is based on.
posted by jaduncan at 2:34 AM on June 13, 2013


I'm assuming this is a joke

Uh, yeah, I just assumed it'd be true. Is it true?
posted by pompomtom at 3:42 AM on June 13, 2013


Just looking at the musical genre database. So much wrong I don't know where to start.
posted by acb at 3:45 AM on June 13, 2013


> Google has announced as of January 1st, 2014, this service will no longer be available.

Really? If so, don't fret. Freebase was launched back in the Web 2.0 era, when companies thought getting your data out was a feature. You can download their data under a CC-BY license.
posted by Monochrome at 4:36 AM on June 13, 2013


Uh, yeah, I just assumed it'd be true. Is it true?

It's not mentioned on the site, updated-twice-in-the-last-week G+ feed, or on the Wikipedia entry. I'm going with no.
posted by jaduncan at 5:15 AM on June 13, 2013


Drill, baby, drill.
posted by Thorzdad at 5:58 AM on June 13, 2013


Uh, yeah, I just assumed it'd be true. Is it true?

I'm fairly certain it was a joke on the fact that Google Reader is shutting down on July 1st, but I think that it says something that everyone seems to assume that it might be true.

The reason I'm fairly certain it's a joke is that I was going to make the exact same joke...
posted by gkhan at 6:08 AM on June 13, 2013 [3 favorites]


I'm not going to use Freebase. I remember what happened when Richard Pryor tried it.
posted by Sys Rq at 6:24 AM on June 13, 2013 [2 favorites]


hal_c_on: "Music- 25 MILLION Facts.
Physics- 55 facts.

This is what happens when the internets takes over.
"

That's cuz there's physics "THEORIES"!
posted by symbioid at 6:36 AM on June 13, 2013 [2 favorites]


Freebase is awesome, but for the life of me, I still cannot figure out how to write queries in FQL.

The data's also lacking in a number of areas (I was hoping that their Movie/TV stuff would be better, but it's pretty lacking at the moment.)

IMO, to curate better data, Google should build some topic-specific front-ends to the database to allow people to view and update the data in a Wiki-Like format. Right now, Freebase's powerful schema imposes a serious learning curve on anybody who needs to use it or contribute data.
posted by schmod at 6:53 AM on June 13, 2013


Wikidata (funded in part by Google)
posted by stbalbach at 7:22 AM on June 13, 2013


Among the bottom categories are Physics, geology, bicycles - all subjects Mefi does well. Also, couldn't see politics or gender sections.
posted by marienbad at 7:37 AM on June 13, 2013


Huh, I'd forgotten all about this. Google bought Metaweb in 2010 for an "undisclosed sum". Given the lack of attention this product has gotten since, I'd thought it was mostly a talent acquisition and the company was for sale cheap. Hard to say though; no purchase price ever leaked. I'd assumed this was part of the Knol effort which got wound down last year?

Danny Sullivan notes that Metaweb is part of Knowledge Graph, Google's name for the feature that shows you sidebar summaries when you search for certain factual triggers like Matt Haughey. So maybe the technology is woven deeper in the search engine and FreeBase is just a sort of awkward UI for accessing it.
posted by Nelson at 9:08 AM on June 13, 2013


Freebase was acquired and morphed internally to become the Knowledge Graph.
posted by zippy at 9:28 AM on June 13, 2013 [1 favorite]


Right now, Freebase's powerful schema imposes a serious learning curve on anybody who needs to use it or contribute data.

Yeah, this is way too complex and imposing for a sizable percentage of the people who could contribute to it. Hell, at the very first step of choosing a username, here's the error message if you choose an invalid one: Invalid value 'Hargrimm'. Values must match the following regular expression: '^[a-z][a-z0-9_]{1,30}[a-z0-9]$'

The very first thing they ask a prospective user to do is interpret a regular expression!
posted by Hargrimm at 11:06 AM on June 13, 2013 [5 favorites]


When I just tried to make the account

d______________________________b

I got a page with a 503 error. In JSON.

{"message":"urlfetch failed: 503","response":{"status":503,"body":"{\n \"error\": {\n \"errors\": [\n {\n \"domain\": \"global\",\n \"reason\": \"backendError\",\n \"message\": \"Backend Error\"\n }\n ],\n \"code\": 503,\n \"message\": \"Backend Error\"\n }\n}\n","headers":{"X-Frame-Options":"SAMEORIGIN", ...
posted by zippy at 12:02 PM on June 13, 2013


Did it really throw a regexp at you? I tried to create an account with a space in the name, and it just gave me an "Invalid user name" error.

That being said, that's a fairly straightforward regexp.
posted by schmod at 9:26 PM on June 13, 2013


that's a fairly straightforward regexp.

"Fairly straightforward" is the academic's "hold my beer and watch this."
posted by zippy at 10:36 PM on June 13, 2013 [2 favorites]


« Older "I've worn the same outfit as my husband for 35...   |   SPOILER Newer »


This thread has been archived and is closed to new comments