Monday, April 23, 2012

Monday Muse Spotlights The Book Genome

Our goal is not to make it faster to find books, exactly,
but to make discovery more enjoyable. . . .
~ Aaron Stanton, CEO, Novel Projects, Inc./BookLamp*

Who knew books have "genes"? Or could be reduced to thousands of data points?

The fascinating Book Genome Project, originally the brainchild of a group of University of Idaho students seeking to determine how books manage to stand out and get read in the overcrowded "bookosphere", subjects the language, theme, and characters of books — digital text is provided by publishers — to computer analysis that yields for each of the three primary metadata a subset of measurements ("StoryDNA", "Language DNA", "Character DNA") that are analyzed and broken down further into individual thematic "ingredients" and variable components until a complete "gene structure" is created. Visually, the plotting of a book's "literary DNA" broadly looks like this:

According to the project site, every book produces 32,162 genomic measurements (data points such as setting, pacing, dialogue) that can be identified, classified, tracked, compared with other books, and studied to produce personalized ratings-based reading recommendations. The idea is to match books' content (not, for example, a reader's purchase history or writer popularity) with readers.

The results obtained with the technically complex algorithms are applied at BookLamp, the project's non-commercial "public face", a book-discovery search engine that helps connect readers directly to the kinds of books they enjoy reading. (See the FAQ section at BookLamp for information about the technology. Also see "Understanding the Book Genome Project" on BookLamp's blog.) A reader receives a list of suggested books to discover by going to BookLamp's Homepage and entering an author's name or book title in the search box, or by browsing by StoryDNA, Author, or Genre (currently, fiction and nonfiction). It costs nothing to use the tool. 

A book has to have been issued by a publisher working with Book Genome to be in BookLamp's database. The project, which frankly admits it needs more books in its database to reach its potential, is actively soliciting publishers' participation to help reach its goal "to make all books discoverable". 

BookLamp on FaceBook and Twitter

Also Of Interest

Edward Nawotka, "Is BookLamp's 'Book Genome Project' the Future of Discovery?", PublishingPerspectives, August 24, 2011

* Sandra Forester, "Boise-Based Startup BookLamp Looks at Books a Different Way", Idaho Statesman, August 16, 2011 (This is a Q&A with concept originator Aaron Stanton.)


Kathleen said...

Neato! I love that this joins the human genome project and the music genome project as a way of thinking about how things connect..and how we, as humans, categorize, etc.

Anonymous said...

what came first; the book or the book genome?