The Wayback Machine - https://web.archive.org/web/20101015090119/http://bldgblog.blogspot.com/2009/01/bldgblog-book-as-series-of-word.html
BLDGBLOG: The BLDGBLOG Book as a Series of Word-Frequency Clouds
close



The BLDGBLOG Book as a Series of Word-Frequency Clouds

I got to playing around with Wordle yesterday morning, and began to make word-frequency clouds for the entire text, captions and all, of The BLDGBLOG Book (check out a few preliminary spreads here, by the way).

BERJAYAYou can toggle the inputs; the following images thus represent everything from the 250 most common words to appear in the book to the 15,000 most common words (I'd be hard-pressed, on the other hand, to believe that there are actually 15,000 different words in the entire book).
Check out the Flickr set for more.

BERJAYABERJAYASo I realize this won't be of interest to everyone, but I think it's pretty amazing, personally, to see an entire manuscript reduced to this.
For instance, is it possible to extrapolate from these things what the original book might have been about? Or could you write a new book based on these word-clouds, which you would then compare back to the original? If you wrote a book based on an existing book's word-clouds, would the word-clouds of that second book be the same? Or could you then write a third book based on the second book's word-clouds – etc. etc.?
And might it be possible for two very different books – say, a popular history of aviation and a new book about climate change science – to generate all but identical word frequencies?

BERJAYAAs I joked on my Twitter feed yesterday, it's like Rorschach literature, literary cobwebs from which you can pick and choose new meanings. "Subterranean time museum." A "living tunnel designed outside." "Artificial concrete."
"Another else begins beneath nature."
"Much water writes landscapes."

BERJAYABERJAYAIf there's something called Wordle, meanwhile, it'd be interesting to see something like Roomle: a program into which you can enter buildings, or clusters of buildings, and the most common room-types come back, organized into hierarchical clouds. You plug all of New York City into the Roomle generator, and find that the most common room of all is...
For a student project, you then construct those Roomle clouds as real buildings, bizarre labyrinths of competing room-types in which a visitor could be lost for days. It'd be a kind of baroque postmodernism, mathematically reached.

(Wordle first spotted via Classic Detritus).

For posts older than a few days,

comments have to be moderated;

this takes a few hours. Thanks!


html is both okay and encouraged.





18 Comments:

Blogger Anna said...

Wow this is so interesting. You have created some amazing architectures. Thumbs up! I am going to try it myself. Wow again, Anna :)

January 30, 2009 8:45 AM  
Blogger mv said...

Have you read Italo Calvino's "If on a winter's night a traveler"? It is, in part, a narrative about a reader trying to find the complete copy of a particular book [in this case, Calvino's "If on a winter's night..." -- meta-literature?].

At any rate, the reader encounters another reader who uses a computer program to read for her. The program generates lists of keywords from a particular book; the reader can then skim the generated list, and have a sense for what the book is about, even without opening to the first page. Similarly, there is a corporation that generates spurious works by any particular author through commonalities within the other works.

Definitely an interesting read, if you ever have the time.

January 30, 2009 9:24 AM  
Blogger mv said...

Forgot to mention -- the spreads looks great!

January 30, 2009 9:27 AM  
Anonymous Mike said...

Geoff, brilliant!

January 30, 2009 11:53 AM  
Anonymous Dave said...

Geoff, is it going to be possible to buy the book in the UK? Does the publisher ship here? Also is the book only available direct from your publisher, or other providers e.g.amazon etc.

Thanks

January 30, 2009 3:41 PM  
Blogger Plinius said...

Great stuff - just tried it on my blog and it looks pretty good.

January 31, 2009 6:26 AM  
Anonymous javier said...

The cover & spread looks great! Give my best to Brett And Scott. From Ithaca,
Jav

January 31, 2009 3:07 PM  
Blogger Kevin Broome said...

Or you could take your Roomle results and stack the rooms so the most repeated are on the bottom level and the rarer rooms are up top thus creating giant pyramids with first floors made up entirely of closets, then bathrooms or bedrooms on the second floors etc. The rooms that made up the pyramid's point would be things like wine cellars, maid's quarters and panic rooms.

January 31, 2009 3:36 PM  
Blogger Speed McQueen said...

Geoff,

I love that "like" is so huge, perhaps because of it's multivalence as both an expression of approbation and as the hinge of a simile. I wonder which use is more frequent. Drew also pointed out that "thus" fared well. Unsurprising as it is the most sophisticated coordinating conjunction. "Actually" appears to be the most profligate adverb, and "new" outstrips all other adjectives. I love it.

January 31, 2009 7:30 PM  
Anonymous Hannah said...

Too add to the love for "small" words, the first thing I noticed was prevalence of the word "perhaps". Fitting, considering speculation is so constitutive of this blog, and is one of my favorite things about it.

February 01, 2009 10:04 AM  
Anonymous G said...

Anybody else wishes for the old cover? You know, the one that has been stuck to the side of the site for a while? it was so clean and different... oh well, i am sure some marketer told you you had to add "the" and "book" and some pretty pictures. too bad.

February 02, 2009 10:30 AM  
Anonymous G said...

ps. Totally buying it anyway! the spreds look great!

February 02, 2009 10:31 AM  
Anonymous David Maisel said...

Any of these examples could make for a fascinating book cover for the BLDGBLOG book...They somehow convey what one will find within the book's pages. Too late for re-design? Then startithinking about the 2nd edition!
Cheers,

David

February 02, 2009 1:47 PM  
Blogger Alexander Jack said...

Shortly after I saw your post, I found this amazing Wordle of past inaugural speeches. http://www.research.ibm.com/visual/inaugurals/

February 02, 2009 6:28 PM  
Blogger kolaslaw said...

<3 wordle! your wordl(es?) came out so much more interesting than anything I've ever thrown in there.

February 02, 2009 9:23 PM  
Blogger Kosmograd said...

Hey, the spreads look great, though not as good as Grilling with Dad or Syrian Lingerie!

Can't wait to read it. With your book, Owen H's Militant Modernism and Adam Greenfield's The City is Here to You to Use, it's going to be a good year for Bloggers Publishing Books (BPBs)!

February 03, 2009 3:25 AM  
Blogger TT said...

Look's very cool!

February 04, 2009 3:11 PM  
Blogger Geoff Manaugh said...

Thanks for the comments! And sorry to be jumping in here only now; it's been a hellaciously busy week. The cover that you see on the Chronicle site, as well as on the site of the designers, is not the final cover - though you do get a very clear idea of where the cover is going from those glimpses.

The book is more or less now with the printer, incidentally, so this thing is finally crawling into the light. Look for more updates, previews, and images from the book over the next few months - and I think you'll like the ultimate result. It's got several hundred images, something like 110,000+ words, new interviews, two original comic strips, five major chapters, dozens of shorter subchapters, a massive reading list, and all sorts of other things - but I'll keep talking about all this stuff as the release date approaches.

Anyway, thanks again for the comments - and sorry it's been a slow week here, with new posts, but I've cleared my desk of a lot of freelance projects and can get back to posting now...

February 07, 2009 11:59 AM  

Post a Comment