Thursday, March 15, 2007

"Wow, data is fun."


There's a nifty post over on the Google Book Search Blog where Google engineer Matthew Gray charts locations mentioned in books on a world map, and then filters by publication date into a series. Gray notes that it picks up the westward expansion of the United States. It picks up some other events too. The Scramble for Africa is noticeable, if from a largely British perspective.

I like the way he closes—"Wow, data is fun." My feelings exactly. It's why LibraryThing has five recommendation algorithms (not counting two I'm hiding). It's why we have a "fun statistics" page that reports on users with whom you share the only two copies of a work on LibraryThing (the much misunderstood Vous et nul autre feature). It's why I'm giddy that LibraryThing has the largest collection of book tags on the web.

But not everyone has query-level access to LibraryThing's data. We need to get more out there, so members and passers-by can play with LibraryThing's increasingly rich dataset as I do. We'll have some news on that front soon.

8 Comments:

Anonymous Anonymous said...

I'd be happy to see a "more robust" Fun Stats page that incorporated data from the Zeitgeist (i.e. "With ### reviews, you are the ###th Most Prolific Reviewer on L.T.", "With #### books, your library is the ###th largest on L.T.", etc.).

Also, having "bell curve" distribution graphs (showing where a user's statistical data fell in regards to the whole of L.T.) for the various statistical categories would be swell!

3/15/2007 1:49 PM  
Anonymous Anonymous said...

Oh, wow, query-level access?? Oh, please! Please-oh-please-oh-please! If you guys haven't got THE most fun sandpile-o-data on the 'net, then I can't imagine who does!

3/15/2007 2:05 PM  
Anonymous Anonymous said...

I just have to mention that one should take the Google data with a grain of salt. The publication dates given by Google Books are completely unreliable, often off by a century or more.

Now if there were a way to connect LibraryThing publication data with Google Books text search capability...

3/15/2007 3:27 PM  
Anonymous Anonymous said...

I've long thought about begging for query level access—at least to my own library.

3/16/2007 10:10 AM  
Blogger Soji Slade said...

Interesting, the website you have linked to, BookSearch leads to an error message. Apparently, your link overloaded the server (well, the people heading over through your link). :) Wish I could have read the "nifty post."

3/16/2007 1:11 PM  
Anonymous Anonymous said...

You're hiding recommendation algorithms? C'mon Tim, information wants to be free!

3/16/2007 1:50 PM  
Blogger Tim said...

The algorithms or the data? The algorithms are basically statistics...

3/16/2007 3:32 PM  
Anonymous Anonymous said...

Some people seem to be polluting the data, though.

Check out the no 1 Zeitgeist author Trish New, check which people have given her top ratings, and check what books out of their very small libraries they've chosen to review...

3/19/2007 3:09 PM  

Post a Comment

<< Home