by Joe Miller,
Director of Development
Last week, Hallie Herold and I had the pleasure of representing the Pocahontas County Library system at our quarterly service area meeting in Lewisburg.
One of the highlights for me was learning that the state provides free access to a service called NoveList.
NoveList is a tool for helping you find books you’ll like.
If you’ve ever used things like Netflix or Spotify, you’ll have a good idea of what NoveList does. You tell it things you like and it’ll recommend other similar kinds of things.
Just search for an author, book or series you like. In your results, you’ll see the option to choose what NoveList calls “read-alikes.” That’s just what it sounds like—a set of things that are similar to a book you’ve enjoyed. I tried it with a few recent favorites. The site immediately generated a list of books that I’d already read and enjoyed—along with a few new ones that are now part of my to-be-read stack.
I find NoveList really fascinating because it sits at the intersection of the two different types of work I do—working with the library and building complex websites. So this week, I’m going to talk a bit about how those things overlap.
Buckle up, because it’s about to get a little nerdy here.
Both libraries and websites are powered by a thing called metadata. This is a fancy term that really just means data about data.
As an example, last week, I checked out a book called “Hippie Homesteaders.” It’s about the history of the back-to-the-land movement in West Virginia in the 1960s and ‘70s. The actual words and photographs in the book are its data.
But there’s a lot of other information about the book. It has a publisher. A publication date. An author. An edition number. An ISBN.
Those things are metadata.
The book also has a lot of more conceptual metadata. It’s a work of nonfiction. It contains a bunch of biographies of people who settled in the state. It’s about some things—West Virginia, artisans, counterculture, migration, rural life.
Libraries have developed a standardized set of subject metadata for nonfiction books. Those sets of subjects allow librarians to shelve similar kinds of books together.
You might remember learning the Dewey Decimal System in school. That’s built on a set of shared subject metadata. You can walk into any community library in the country and find all the books about mathematics in the section with the number 510. Really large libraries—those in big cities or attached to universities—use a different system (the Library of Congress Classification system), but it still operates from the same basic principles.
That all means that organizing books in libraries requires two things. Someone has to read each new book and figure out what it’s about—that is, which of the standard categories does the book fall under. And then someone has to add all that information into our catalog.
The publisher generally handles the first of those tasks. Our team of librarians add metadata for books into our catalog.
Websites also rely on metadata. In fact, when we first started building websites, we borrowed a lot of ideas from the library sciences.
When you go onto a website to buy something—suppose you’re looking for a new pair of shoes—you can filter by things like size and color and style and material. Those are all types of metadata.
Services like Netflix and Spotify work the same way. There are dozens and dozens of hidden metadata fields for each show, movie, song or album.
They can then recommend new things by looking at the metadata attached to the things you play and then finding other things with a similar set of characteristics. The more metadata fields that match, the higher the recommendation appears. And the more you watch/listen to things, the better the algorithms get because they have more metadata to match.
That’s how Netflix ends up with weirdly specific categories (e.g., showbiz mo-vies based on real life or tearjerkers from the 1970s).
NoveList is really cool because it’s actually part library and part website.
See, while we have a really sophisticated system for classifying and organizing nonfiction in libraries, our system for organizing fiction is much simpler. They’re alphabetical, based on the author’s last name.
We sort by age, and sometimes we separate out genre fiction (e.g., romance, science fiction, westerns, thrillers) or by type (graphic novel).
But mostly it’s just last name. You might find Ste-phen King’s contemporary horror novels shelved right next to one of Charles Kingsley’s 19th-century no-vels extolling the virtues of Christian socialism.
The people behind NoveList are creating new sets of metadata for fiction.
Some of that is around writing style (e.g. fast-paced, quirky, cynical humor) and some is around themes (e.g. friends-to-lovers, manifest destiny) and some around sub-genres (e.g. military science fiction, romantasy).
Combine enough of those things, and you can start to get some uncannily accurate recommendations.
But, look, you don’t have to know how NoveList works to reap the benefits. Stop by any local branch and we can get you access. Once you have a username and password, you can use it from anywhere.
It’s a cool way to find your new favorite author.
