Learn More: Tags and Tag Clouds

December 3, 2007

[This is one in a series of self-paced discovery exercises for library staff venturing into the social web.]

We explored social bookmarking last week and briefly mentioned tags. Let’s look at tags — or folksonomies — a bit more closely now.

Formal cataloging involves experts who assign subject headings based on approved lists and hierarchical taxonomies. It’s an organized and presumably consistent way of storing and retrieving materials and information. But there are many ways of classifying things. I might look at a photograph of a lake in winter and think of the words lake, frozen, and cold. You might prefer lake, winter, and ice. We often rely on experts to iron out our differences. Unless we think like the expert who cataloged the item, however, we still might not find what we’re looking for.

Now suppose everyone participated in cataloging. You might imagine people attaching notes to everything (much like the two girls in a delightful video from Harris County Public Library). With folksonomies, you tag things the way you want, and I tag them the way I want. If enough people participate, we might collectively describe any content with greater depth, color, and nuance than formal subject headings.

Consider J. D. Salinger’s Catcher in the Rye. The book’s formal cataloging in WorldCat shows:

Caulfield, Holden (Fictitious character), Runaway teenagers – Fiction, New York (N.Y.) – Fiction

The collective input of 1000+ people tagging the book in LibraryThing offers all these words:

Because content creators and users generate the tags, folksonomies often present information that the average person finds most relevant. Visual representations of tags (like the one from LibraryThing shown above) are called tag clouds and may give us even more information because they display frequently-tagged words in larger fonts. More people thought those words were important.

Visitors to social networking sites tag content for their own reasons. Each person adds words relevant to him/her, and the aggregation has a powerful organizing effect. If you’d like to know more about tagging and organizing digital information, I recommend Everything is Miscellaneous [LibraryThing/WorldCat], a very readable book by David Weinberger.

Meaning for libraries

Libraries cannot possibly classify everything. Even the enormous Library of Congress (with several hundred catalogers sorting through about 7,000 new books each day) could not handle the volume of Flickr, the photo-sharing website we discussed a few weeks ago. Flickr receives a few thousand photographs every minute. Tagging makes possible the organization of all those pictures.

YouTube members tag videos; Delicious users tag websites; LibraryThing members tag books; bloggers tag their posts. Tagging is everywhere on the social web.

It’s vital for library staff to understand tags. They should know the strengths: that tags accumulate quickly and offer a wide range of keywords. They should be aware of the weaknesses, too: tags can be ineffective when very few people are involved and offer inconsistent word choices (examples: Should we search car, cars, or auto?; Which meaning of the word spring was intended?).

Libraries might also explore ways to incorporate tags into software where appropriate. Some libraries have or soon will have tags and tag clouds in their catalogs. My own library has tags on its staff intranet site.

Learn more by participating

You’ve read enough by now, so this week’s activity can be done in short exploratory sessions of mouse clicking.

  • Explore your Flickr tag cloud. Assuming you created an account for an earlier lesson, log into your Flickr account and go to the “Your Tags” option in the “You” menu. This cloud was assembled with the tags you added to your own photographs. Common themes among photos shine through. Click any word to find all photographs marked with that tag.
  • Look at Flickr’s communal tag cloud. This will call up other Flickr members’ photos. It’s great for browsing, but there’s a search box, too, in case the word you’re looking for isn’t there.
  • Explore tags in your Delicious account. (That was a previous lesson, too.) Your tags appear in a clickable list or cloud on the right side of the screen. You may also click the “saved by X other people” link shown for any particular bookmark and see the collected tags from everyone else who saved that bookmark. Click one that interests you and see what other sites come up.
  • Add more tags to your Flickr photos and Delicious bookmarks. It will make them easier to find and others to stumble upon.
  • Explore Dave Pattern’s biblioblogosphere tag cloud. Pattern’s software regularly scans library blogs and updates the communal tag cloud. A glance at the cloud can give you a sense of the topics librarians around the world are currently chattering about. Clicking any word in the cloud takes you to those blogs doing the chattering.
  • Create a text cloud. Just for fun, copy the text of a long memo or email into TagCrowd and see the cloud it produces. The longer the text, the more likely the prominent words will be meaningful.

Mashing LibraryThing and Polaris

October 8, 2007

The other day I mentioned computer code with which Polaris was toying to add more “funability” (Tim Spalding’s word) to the library catalog. Here’s how it would work:


Let’s say you bring up Lois Lowry’s The Giver. The first part of that bibliographic record is straight out of the library’s PAC, but nicely tucked below is a tag cloud piped in from LibraryThing‘s database of 25 million tags. Catalog users would get great visual cues about the book from the few thousand readers who tagged their own copy of the book in LibraryThing.

The words are not fixed LC subject headings. They were entered by readers of the book. That means the words will more likely relate to the readers browsing your catalog. There’s real benefit to that. (LC subject headings are still available in the PAC’s detailed view, mind you. For The Giver, they lead off with “Euphemism” and “Euthanasia”. Yawn.)

But the fun doesn’t stop there…


Click one of the tags and some LibraryThing javascript (inspired by LightBox) generates a long list of books that best match that particular tag. Who made that list? Everyone who tagged their own books. It’s an example of social website collaboration.  Tim Spalding has done such a remarkable job finding relationships in all his data crunching that the results are eerily good. What’s more, the titles returned from LibraryThing seamlessly filter out books not in your library’s catalog. Your patrons won’t be bogged down with stuff they can’t readily get their hands on. They can also continue clicking and browsing.

Tim described this concept months ago and has been able to tack it onto the catalogs of a few individual libraries already. I previously mentioned Danbury as the first example but, as far as I know, Bryan Rubenau at Polaris is the first to write code toward fully implementing this stuff into the guts of an ILS software package. His preliminary code rocks. If it goes into a future upgrade*, catalog users might soon search the reader-contributed tags (like “chick lit”) just as easily as they search titles, authors, and subject headings.

Impressed? Join the club.


*At the moment, Polaris is looking into the specific requirements for integration and asking its customers if they’re interested. There’s no formal partnership between LibraryThing and Polaris.


Snippets of PUG

October 6, 2007

Conferences dispense so many topics that it’s hard to pluck just one headline from the annual Polaris Users’ Group (PUG) Conference that wrapped up in Syracuse, NY today. I’ll mention a few snippets now and elaborate later.

Tim Spalding, founder of LibraryThing, delivered a good — often hilarious — keynote address yesterday urging libraries to make their catalogs more fun, and more willing to use data tools that discover book relationships not tied to standard subject headings. His Death Star metaphor for OCLC might be a tad over the top, but I’ll gladly follow him down the road of tags and tag clouds. He’s done amazing things with his database of nearly 25 million member-generated tags. (More on that in a future post.)

As a fan of social networking tools, my favorite news of the week came just after the keynote: Polaris programmers have written code exploring a connection with LibraryThing. About six months ago Tim started talking up the possibility of integrating catalogs with LibraryThing tag clouds and related books lists. (I remember this clearly because we tried to engrave our library’s name on his list the following day.) The Danbury Library has it working already. It’s pretty slick. In recent weeks, Polaris programmers were intrigued enough to start writing code to its ILS software to maximize all the bells and whistles that would come streaming in from LibraryThing. There’s no formal partnership between the two, mind you, and no certainty that the coding will go into a new release, but there’s enough written code that Polaris gave us a working demo. I’m a bit partial to the toolkit, mind you, but tag clouds and tag-generated book suggestions in the Polaris ILS looks fantastic. I’ll post screen shots in LibraryStream as soon as I get them.

I was also intrigued to learn more details about the Dewey-free Perry Branch in Maricopa County (AZ). They made news around the world earlier this year when they opened with a design modeled on contemporary bookstores. Materials are arranged by book industry labels (like “Pets” and “Cooking”) rather than Dewey Decimals (like 636 or 641). Cindy Kolaczynski, Maricopa’s Deputy Director, gave background and a progress report at PUG this morning. I’ve got my doubts about going entirely bookstore-based, but love the spirit of experimentation at Maricopa. If no one tries things like this, we’re all just flapping about theory. I’ll come back to this topic, too.

I gave a presentation at this year’s conference — “Training 2.0: How to expose and inspire your staff to the social web” — but I hope that this blog will give you the gist of that over time.

Finally, PUG reminded me the value of face-to-face, meet-in-the-hallway connections that happen so often at conferences. To the many new people I met in Syracuse: It’s a pleasure. To those I had met before: Our conversations seemed to simply pick up where we had left off. To those of you going to other conferences: Eat them up.