Charleston 2014 – Crowd Sourcing of Library Services

Speaker: John Dove, Credo Reference

Lots of words about what crowd sourcing is and why we should care. This is why I’m not a scholar. Just get to the point and don’t spend so much time convincing me that the point is the point.

 

Speaker: Tim Spalding, LibraryThing

It’s a personal cataloging tool that becomes social with more people doing it. Personal cataloging is the basis, and it was started with the idea that it would only be that.

Users can add tags to categories their books, and there are over 112 million tags from users. Users can add cover images for their own books, creating a vast collection of book covers.

The next level of engagement is exhibitionism and voyeurism, followed by self-expression via reviews. Reviews happen after a person reads the book, not when they are looking it up in the library catalog.

Users can add their own series information, including sub-series, which is often more information than what librarians are able to add. Other common knowledge content includes characters, author information, etc. Members also manage the “authority control” — FRBRization, author disambiguation, tag disambiguation.

Policing (get off my lawn) and helping (here’s how to be on my lawn) — dealing with spam, trolls, etc. and also assisting newer users via forums.

The final level of engagement comes with collaborative cataloging of books by dead people or that have shown up in mass media (i.e. Dr. Horrible’s Sing Along Blog).

Lessons: secure the bottom of the ladder, build it rung by rung or at least think about it that way, and finally, crowd sourcing is not a feature. It’s not about what you get, it’s about what you give.

 

Speaker: Scott Johnson, ChiliFresh

If the internet in the 70s and 80s, waterbeds would probably have not reached the 20% penetration due to online reviews.

The wisdom of crowd source information is also the madness.

Rather than having a closed database of reviews from local patrons only, they have a collaborative database of reviews from library users across the world that local libraries can choose to participate in it or not. The reviews themselves are written by patrons, but they are moderated by librarians.

 

Speaker: Ilana Stonebraker, Purdue University

How is my library like the Vlog Brothers 54 jokes video? There is a huge network and community doing important things that are not visible in just that video. We are icebergs.

Most reference questions are lower-level, even online questions. What’s supposed to happen is at a much higher level, but the reality is that isn’t most of what happens. The traditional reference service model also assumes that the librarian is the only one who can give the answer. For example, sometimes students who have had a similar problem and found a solution can help each other.

CrowdAsk is similar to StackOverflow for gamification and badging. It’s open source on Git Hub. You can ask a question, and assign a bounty using your points to get a faster answer. They use it in lower level courses to allow the students to work together. Users can vote on answers and questions. Students who are really good at answering each other’s questions gain more power/authority in the system.

There is a good level of participation so far, and there are quite a number of lurkers, with the average time spent at over 6 minutes. They did some usability tests and found that often the motivation is reciprocity — they were helped and they want to help others.

The goal is to create a sustainable user engagement and community involvement as a part of the library’s website, not just to triage late-night reference questions.

ER&L 2010: Adventures at the Article Level

Speaker: Jamene Brooks-Kieffer

Article level, for those familiar with link resolvers, means the best link type to give to users. The article is the object of pursuit, and the library and the user collaborate on identifying it, locating it, and acquiring it.

In 1980, the only good article-level identification was the Medline ID. Users would need to go through a qualified Medline search to track down relevant articles, and the library would need the article level identifier to make a fast request from another library. Today, the user can search Medline on their own; use the OpenURL linking to get to the full text, print, or ILL request; and obtain the article from the source or ILL. Unlike in 1980, the user no longer needs to find the journal first to get to the article. Also, the librarian’s role is more in providing relevant metadata maintenance to give the user the tools to locate the articles themselves.

In thirty years, the library has moved from being a partner with the user in pursuit of the article to being the magician behind the curtain. Our magic is made possible by the technology we know but that our users do not know.

Unique identifiers solve the problem of making sure that you are retrieving the correct article. CrossRef can link to specific instances of items, but not necessarily the one the user has access to. The link resolver will use that DOI to find other instances of the article available to users of the library. Easy user authentication at the point of need is the final key to implementing article-level services.

One of the library’s biggest roles is facilitating access. It’s not as simple as setting up a link resolver – it must be maintained or the system will break down. Also, document delivery service provides an opportunity to generate goodwill between libraries and users. The next step is supporting the users preferred interface, through tools like LibX, Papers, Google Scholar link resolver integration, and mobile devices. The latter is the most difficult because much of the content is coming from outside service providers and the institutional support for developing applications or web interfaces.

We also need to consider how we deliver the articles users need. We need to evolve our acquisitions process. We need to be ready for article-level usage data, so we need to stop thinking about it as a single-institutional data problem. Aggregated data will help spot trends. Perhaps we could look at the ebook pay-as-you-use model for article-level acquisitions as well?

PIRUS & PIRUS 2 are projects to develop COUNTER-compliant article usage data for all article-hosting entities (both traditional publishers and institutional repositories). Projects like MESUR will inform these kinds of ventures.

Libraries need to be working on recommendation services. Amazon and Netflix are not flukes. Demand, adopt, and promote recommendation tools like bX or LibraryThing for Libraries.

Users are going beyond locating and acquiring the article to storing, discussing, and synthesizing the information. The library could facilitate that. We need something that lets the user connect with others, store articles, and review recommendations that the system provides. We have the technology (magic) to make it available right now: data storage, cloud applications, targeted recommendations, social networks, and pay-per-download.

How do we get there? Cover the basics of identify>locate>acquire. Demand tools that offer services beyond that, or sponsor the creation of desired tools and services. We also need to stay informed of relevant standards and recommendations.

Publishers will need to be a part of this conversation as well, of course. They need to develop models that allow us to retain access to purchased articles. If we are buying on the article level, what incentive is there to have a journal in the first place?

For tenure and promotion purposes, we need to start looking more at the impact factor of the article, not so much the journal-level impact. PLOS provides individual article metrics.

IL2009: Mashups for Library Data

Speakers: Nicole Engard

Mashups are easy ways to provide better services for our patrons. They add value to our websites and catalogs. They promote our services in the places our patrons frequent. And, it’s a learning experience.

We need to ask our vendors for APIs. We’re putting data into our systems, so we should be able to get it out. Take that data and mash it up with popular web services using RSS feeds.

Yahoo Pipes allows you to pull in many sources of data and mix it up to create something new with a clean, flow chart like interface. Don’t give up after your first try. Jody Fagan wrote an article in Computers in Libraries that inspired Engard to go back and try again.

Reading Radar takes the NYT Bestseller lists and merges it with data from Amazon to display more than just sales information (ratings, summaries, etc.). You could do that, but instead of having users go buy the book, link it to your library catalog. The New York Times has opened up a tremendous amount of content via APIs.

Bike Tours in CA is a mashup of Google Maps and ride data. Trulia, Zillow, and HousingMaps use a variety of sources to map real estate information. This We Know pulls in all sorts of government data about a location. Find more mashups at ProgrammableWeb.

What mashups should libraries be doing? First off, if you have multiple branches, create a Google Maps mashup of library locations. Share images of your collection on Flickr and pull that into your website (see Access Ceramics), letting Flickr do the heavy lifting of resizing the images and pulling content out via machine tags. Delicious provides many options for creating dynamically updating lists with code snippets to embed them in your website.

OPAC mashups require APIs, preferably those that can generate JavaScript, and finally you’ll need a programmer if you can’t get the information out in a way you can easily use it. LexisNexis Academic, WorldCat, and LibraryThing all have APIs you can use.

Ideas from Librarians: Mashup travel data from circulation data and various travel sources to provide better patron services. Grab MARC location data to plot information on a map. Pull data about media collection and combine it with IMDB and other resources. Subject RSS feeds from all resources for current articles (could do that already with a collection of journals with RSS feeds and Yahoo Pipes).

Links and more at her book website.

LITA 2008: What is "Social Cataloging" and Why Should You Care?

“Having games in the library strikes me as being like having bocce in the frat house.”

Speaker: Tim Spalding, Founder of LibraryThing

“I have no practical advice for you, but I have inspiration and screen shots.” Such as, images from Dr. Horrible’s Sing-Along Blog and book pile photo submissions.

Social cataloging does not need to be defined. LibraryThing is a good example of social cataloging, but it’s not the only resource out there like that. (LibraryThing is now larger than the Library of Congress.) Good Reads focuses more on the social aspects, and Shelfari is being revived by Amazon. There are other sites like CiteULike and Last.fm that do social cataloging of things other than books.

Social cataloging explores the socialization. LibraryThing embraces the social and the digital because there is no physical aspect (except for what you have in your own collection).

Social cataloging ladder:

  • personal cataloging – your stuff
  • exhibitionism, voyeurism – about you and your stuff
  • self expression – book pile photos, reviews
  • implicit social cataloging – tag clouds on books that incorporate data from all owners, recommendations, connect with other owners of more obscure books
  • social networking – “friends” lists, users who share your books, groups
  • sharing – book covers of different editions, author photos
  • explicit social cataloging – work-level records (any title you would agree on at a cocktail party) for both books and authors, series data
  • collaborative cataloging – building the catalogs of famous dead people, developing an open-source alternative to Dewey

Regarding why Spalding felt it necessary to pull data from libraries and not just Amazon, he says, “Once you are over the age of 30 and you are not a Philistine, you have books that Amazon is not currently selling.”

Interesting factoid about how things are tagged on LibraryThing: LGBT and GLBT tags have two completely different lists of books.

Traditional cataloging is based on the physical form of cataloging with cards. It was too difficult to change subjects or to add weight to particular subjects because you couldn’t do that with physical cards. We need to get away from this now that we have all the flexibility of digital cataloging. Digital cataloging is social cataloging.

LibraryThing users are doing about 1,000 work combinations per day! Voluntarily! Experts on book topics are the ones pulling the data together, not experts on cataloging.

LibraryThing members figured out what books are on Dr. Horrible’s shelf based on a fuzzy still from the video. And then the guy who lives in the apartment where it was filmed corrected the editions listed.

There are many non-librarians who are passionate about books and classification. People care about libraries and library data.

On the other hand, we suck. Our catalogs are fundamentally not open to the web because our pages are often session-specific and not friendly to index spiders. Worldcat.org is getting fewer visitors, whereas Dogster.com is getting more.

Library 2.0 is in danger. Libraries are concentrating on what they can do, not what they can do best. We don’t need to have blogs or pages on Facebook. “Having games in the library strikes me as being like having bocce in the frat house.”

Do not pay anyone for Library 2.0 stuff. Do it yourself. OCLC is not yourself.

Or, pay Spalding for his 2.0 enhancements (LibraryThing for Libraries).

Social cataloging is about the catalog, about what you can do right now, about passion, and about giving (not taking).

thing 11: LibraryThing

I have had a LibraryThing account since mid-October 2005. Most of my collection is in there and tagged, and I’ve even started keeping books in my catalog that I no longer own (appropriately tagged, of course), just so I can keep a record of what I have had at some point.

If you look on my blog, you’ll see that I am using the LibraryThing widget to display a random book from my catalog. This changes every time the page loads, and sometimes I am surprised to see what is there. As I’ve noted several times in the past, I have more books in my house that I have not read than those which I have read.

If you’re new to LibraryThing and you have a large collection of books that are new enough to have barcodes printed on them, I recommend you purchase a CueCat scanner. It will speed up the process of getting everything in, and then you can take more time to tag, make notes, or do whatever else you may want to do to tweak your library to suit your needs.

What I have not done yet is to make use of the Recommendations, mostly because of the aforementioned over-abundance of reading material in my possession. Also, I’ve already read many of the books listed or they are already on my wishlists. Eventually, I plan to import my book wishlists into LibraryThing. I am doing that with my music collection on RateYourMusic, and I can see the value of having all that together in one place.