hangingtogether.org

One of the best parts of my job is working with OCLC Research Library Partner staff on working groups. In this case we never got together face-to-face, but managed to put together a pretty good report, Starting the Conversation: University-wide Research Data Management Policy. Though we started out with a conference call, all the work took place via email and shared documents. The working group consisted of:
Dan Tsang, chair — University of California, Irvine
Anna Clements — University of St. Andrews
Joy Davidson — DCC, University of Glasgow
Mike Furlough — Pennsylvania State University
Amy Nurnberger — Columbia University
Sally Rumsey — University of Oxford
Anna Shadbolt — University of Melbourne
Claire Stewart — Northwestern University
Beth Warner — Ohio State University
Perry Willett — California Digital Library
I supplied the bones, they filled in some of the sections, and I polished it up.
Working in OCLC Research, we try to stay on top of the literature and we hear a lot about application, but there’s nothing like being in the thick of it, so it’s really great to have the expert input of those actually working in research libraries.

As part of my work to reveal exactly how the MARC standard has been used over the several decades it has existed (available at “MARC Usage in WorldCat”), I’ve always wanted to produce visualizations of the data. Recently, with essential help from my colleagues JD Shipengrover and Jeremy Browning, I was able to do exactly that.

After trying various graphical depictions of the data, we finally settled on an interactive “starburst” view of the data. The initial view provides a high-level summary of how often various tags have been used within particular formats. The interactive part allows you to “drill down” a level into a more detailed view.

We are providing two views of the data: from the point of view of the formats being described (that is, the top-level is comprised of the various formats — books, journals, etc.), and from the point of view of the tags (that is, the top-level is comprised of the various MARC tags).

If you have any ideas about a visualization you would like to see, let me know.

During a very hectic, very interesting week visiting research libraries in Japan last week I had the good fortune to tour the new (April 2013) Learning Commons at Doshisha University. It is not a library-managed facility but the library helps to staff it along with other Student Support Services staff. The facility itself is as good an implementation as I’ve seen anywhere including the new facilities at North Carolina State University’s new library.

The Commons itself is a multi-story structure constructed adjacent to the library and connected to the library at various levels. As a consequence students can move very freely from the collections and quiet of the traditional library to the group study, presentation, production and technology areas of the learning commons. There are plenty of visible but unobtrusive staff available to the students. People in red jackets offer technology support, in blue jackets peer instruction and guidance, in yellow you get media production and on each floor a desk staffed by a librarian.

There are no fixed furnishings in the entire facility. Everything can be moved. As an experiment they left one group study space with two tables without rollers. That space is the most infrequently used in the building. I was impressed with the energy of the staff and the enthusiasm of the students. The location of the facility bordering on one of the busiest streets in Kyoto purposely serves to advertise the learning environment of this private university. The big study and computing rooms are lined up along picture windows that face out onto this boulevard ensuring that Kyoto citizens know that Doshisha is a good place to learn.

Check out some photos taken during my walk-through in this Flickr set. Look for the Global Village sign that designates an area where no Japanese is to be spoken.

P.S. After the original post my colleagues at Doshisha advised me that an English language version of their Learning Commons brochure is available (.pdf).

Infoboxes for a long time were Wikipedias’ way of storing data, and Wikidata is set to replace that techonlogy, with added bonuses like inter-language sharing. To get to that promise one first step is for Infoboxes to be harvested into Wikidata. I have started by harvesting Infobox Book in the 9 biggest Wikipedia languages that share the template: English, Italian, French, Spanish, Russian, Polish, Portugese, Swedish, and Japanese.

The point of harvesting Infobox Book specifically is that the Wikidata citation guidelines for books specify that the Library FRBR concept should be used, so I wanted to build out infrastructure to that end. FRBR is about describing Bilbliographic record at many different levels and here’s an example of what this kind of citation would look like in Wikidata:

• Work item: CRC Handbook of Chemistry and Physics (Q904273), edition item: 92nd edition of CRC Handbook of Chemistry and Physics (Q11927173), used to source: ethanol (Q153), statement CAS registry number (P231)

With that in mind lets have a look at the data. O ur entry point is the set of Wikipedia pages that use Infobox Book -transclusion in Wikpedia parlance – in the 9 aforementioned languages. This measure is only an approximation and does not completely reflect how many Wikipedia topics are about books in a language for three reasons. The first is that the conception of a what is a book is not strictly enforced on Wikipedia.  An article could be about a physical item or an amorphous work idea,  or even sometimes the inclusion of an infobox book template is only a nod to a book like French article on this racing pigeon.  The second is that not all articles about a book necessarily contain a transclusion to Infobox Book. And thirdly some specialised Infobox Books have developed and are used instead, like Infobox Doctor Who Book.

In this next chart we look at the total Infobox Book transclusions, the total articles of a language, and the ratio between the two. Despite large variation in absolute numbers, the percentage of Books Articles in a Wikipedia is somewhere beteween .1-1% of all articles. Italians affirm themselves as the most bilbliophilic. We’ll also see later on about how their practice of labelling genre differs from the others.

In each Infobox I crawled for the most used properties across all languages and whose values were either string identifiers or links to other Wikipedia pages. When a value is a link to another Wikipedia page, for instance a link to the page of the author, that is useful because when harvested Wikidata can store the author property as a link to another Wikidata item. This is desirable as in Wikidata we seek to build a Wiki of relations.

Here is a graph of the properties that found, which were added to Wikidatak, and which were already in the database.

So as you can see there are now over 30,000 relations between books and their authors and illustrators in Wikidata, as well as the original language and genres of the books. In addition knowing which book is which from a disambiguation perspective is made easier by the inclusion of over 50,000 identifiers.

One difficulty that was encountered was that even though ISBNs are recorded in Infobox Book, the type of ISBN – 10 or 13 – was not discriminated. Wikidata does however discriminate, and so as I was sorting these ISBNs I thought it would be sage to also verify them. OCLC runs an API called  xID for this very purpose. While using xID it also struck me that the OCLC control number could be returned for a given ISBN. As Wikidata is rapidly evolving into a hub of identifiers, I included those in pushing to Wikidata. During this harvest then I also inserted an additional 10,117 OCNs (not pictured above).

As I mentioned It’s not just boring, nameless identifiers that we want to eventually integrate into all the Wikipedia pages by Wikidata. I inspected genre data as well to see how much cross-cultural benefit we’d receive by doing these sorts of harvests.  Below are the Top 10 genres found in Infobox Book by each language. The text shown are the English Labels of the Wikidata Items of links found in each local Infobox. I’ve also outlined those genres which are unique. So you can see that Swedes care a bit more about the choir books and the Japanese have a bent towards police drama.

What first jumped out at me is how inconsistently the idea of genre is used. In some ways its used to describe the content’s emotion and focus, like “science fiction” or “horror”. Other times its used to describe form like “novel”. In fact only the Italians really are very consistent as their top ten, albeit discusses form in “novel”, “essay”, “short story”, “poetry”, “anthology”, “autobiography”, “novella”, “dialogue”, and “poem”.

Another problem between languages is that the genres mismatch often because they are pointing to only slightly different articles. That is we see appearances from the Wikidata items for “fantasy”, “fantasy literature”, “Fantastique”, and “high fantasy”. (By the way you can draw your own conclusions about the demographics of Wikipedia editors when this much fantasy lit pervades the results.)

A conclusion that can be drawn from all this is that there is still some work to be done on negotiating cultural differences on Wikidata. Wikidata has made a lot of connections between Wikipedia articles in different languages, but not all of those merges are clean. The French conflate a pigeon and a book about a pigeon, and its linked to languages that discuss only the pigeon. Meanwhile how how the Italians interpret “genre” is a different, not necessarily incompatible, notion to others. There are some discussions still to be had probably before Infoboxes completely switch over to using Wikidata data, but we are at least one step closer to that goal.

That was the topic discussed recently by OCLC Research Library Partners metadata managers. It was initiated by Jonathan LeBreton of Temple, who noted the questions staff raised when describing voluminous image collections such as: Do we share the metadata even if it would swamp results? What context can be provided economically? What are others doing both in terms of data schemas and where the metadata is shared?

The discussion revolved around these themes:

Challenges in addressing the sheer volume of digital materials.  Managers are making decisions based on staffing, subject expertise, collection’s importance and funding. It was suggested that some metadata could be extracted from the technical metadata, such as dates and location. We discussed the possibility of crowd-sourcing metadata creation, although experience to date is that a few volunteers are responsible for most contributions, and the successful examples tend to be for transcription, editing OCR’d text, and categorizations. (The At a Glance: Sites that Support Social Metadata chart indicates the ones that enhance data either through improved description or subject access.) The context must matter to people for them to volunteer their efforts. (See the OCLC Report, Social Metadata for Libraries, Archives and Museums: Executive Summary.) With the anticipated increase of born-digital and other digitized materials, there’s a greater need for batch and bulk processing.

Grappling with born-digital materials.  Libraries are receiving the digital equivalents of personal papers and using the Forensic Toolkit to “process” these digital collections.  Preservation and rights management, in addition to description, are important components and no commercially available system yet addresses these needs. The Association of Research Libraries is working with the Society of American Archivists to customize its Digital Archives Specialist (DAS) Program to develop the requisite skills for managing born-digital for ARL library staff. OCLC Research has produced several reports in conjunction with its Demystifying Born Digital program of work.

Concerns about “siloization”, or proliferation of “boutique” collections, using different metadata schema. Metadata is being created in different native systems within an institution, metadata that is often not loaded into a central catalog or even accessible in the local discovery layer. User-created metadata in institutional repositories may be OAI harvested by OCLC and thus may appear in WorldCat even if not visible in the institution’s local discovery tool. Managers grapple with whether to spend resources on updating such metadata before it is exposed for harvesting.  Another challenge is deciding what to include in which discovery layer, and what should be silo’d.  The numerous repositories within an institution can result in complex metadata flows for discovery, as illustrated by UC San Diego’s Prezi diagram. Some institutions map their various metadata schema to MODS (Metadata Object Description Schema), but all non-MARC metadata is converted to MARC when loaded into WorldCat.

What are the “essential elements” to provide access across collections? We posited that librarians have been discussing “core” or “essential” metadata elements for decades, starting with Dublin Core and the Program for Cooperative Cataloging’s “BIBCO Standard Record”. Librarians have been entering metadata for the system it was designed for, but then ultimately the data moves to another system later.  Library metadata is no longer confined to a single system: it may be exposed to search engines and viewed with lots of non-library metadata.

The Library of Congress’ Bibliographic Framework Initiative  portends a future where all metadata will be “non-MARC” and we will rely more on linked data URIs in place of metadata text strings.  How can we use the promise of that future to get to where we need to be?

Contributors to the Virtual International Authority File (VIAF) have generally been national libraries and other library agencies.  We have just loaded into VIAF the first set of personal names from a scholarly resource, the Perseus Catalog hosted by Tufts University, an OCLC Research Library Partner. The Perseus Catalog aims to provide access to at least one online edition of every major Latin and Greek author from antiquity to 600 CE. Adding the Greek, Arabic and other script forms of names in the Perseus Catalog enrich existing VIAF clusters that previously lacked them.

This addition represents a milestone in our Scholars’ Contributions to VIAF activity. We anticipate mutual benefits from our collaboration with scholars. Scholars benefit from using VIAF URIs as persistent identifiers for the names in their own databases, linked data applications and scholarly discourse to disambiguate names in multinational collaborations and using VIAF as a means to disseminate scholarly research on names beyond scholars’ own communities. Both scholarly societies and libraries benefit from enriching VIAF with name authority data which would not otherwise be contributed by national libraries.

As noted in an earlier blog post, Irreconcilable differences? Name authority control & humanities scholarship,  OCLC Research discovered key issues important to scholars that didn’t mesh well with library practices represented in name authority files due to differences in intended audiences, disciplinary norms and metadata needs. However, if scholars do use the Library of Congress’ Metadata Authority Description schema, or MADS, as the Perseus Catalog does, we can add their files to VIAF much more easily.

Adding these scholarly files can demonstrate the benefits of tapping scholarly expertise to enhance and add to name authorities represented in VIAF. We have already seen the number of “alternate name forms” associated with VIAF clusters that include the Perseus Catalog’s contributions increase, with scripts not yet represented. We look forward to more such enhancements from other scholars’ contributions.

Number of institutions with WorldCat holdings for Arabic-language resources

A differentiating feature of WorldCat is that it includes more than two billion holdings of libraries from around the world. My colleague Roy Tennant recently generated statistics on the Arabic-language resources described in WorldCat records. I was struck by the dispersal of the holdings of those materials, as shown in the map above.

Big caveat: Many strong Arabic-language collections are under-represented or not represented at all in WorldCat. Even so, we can see at a glance that Arabic-language materials are collected by institutions in countries far away from the counties of origin, even where Arabic is not widely spoken. Scholarship is international. We could produce similar maps for other language materials.

My colleague Brian Lavoie’s report earlier this year, Not Scotch, but Rum: The Scope and Diffusion of the Scottish Presence in the Published Record, describes in detail his analysis of holdings for materials published in Scotland, by Scottish people, and about Scotland. It concludes, “Most holdings of materials in the Scottish national presence are by institutions outside Scotland, which reminds us that a national presence in the published record may be primarily manifested outside the home country’s borders.”  

Great works are translated—the cream of the world’s cultural and knowledge heritage is shared by being translated. And many of them are represented by bibliographic records in WorldCat.

A group of us working on Multilingual WorldCat projects have been focusing on datamining WorldCat for works and all translations associated with them, identifying the translator for each translation. We plan to generate “uniform title” and “expression” records (the translations) and contribute them to the Virtual International Authority File (VIAF).

We currently have roughly 15 million personal name “clusters” in VIAF, the 26 million personal name authority records contributed by 35 agencies that represented the same person. These are not just creators of works, but also people that have had works written about them and sometimes a translator.

My colleague Jenny Toves has identified about 1 million persons in WorldCat who are associated with bibliographic records in more than one language, or roughly 7% of the people represented in VIAF.  The breakdown:

• 624K names are associated with titles in only 2 languages
• 283K names are associated with titles in 3 to 9 languages
• 7K names are associated with titles in 10 or more languages

Persons with titles in multiple languages

My colleague JD Shipengrover created the accompanying graphic.

As recently announced by Terry Reese, his program MARCEdit now includes a great set of new capabilities for users of WorldCat. Recently made possible by the release of the WorldCat Metadata API from OCLC, here are just a few of the things you can do directly from MARCEdit:

• Set Batch Holdings in OCLC.
• Batch upload/edit records into WorldCat.
• Search WorldCat directly from within MARCEdit.

This is just the kind of integration that our web services now make available for software of all kinds. By providing an application program interface (API) that enables not just search and display of records, but also updating and creating records, we are exposing the full range of WorldCat metadata capabilities to virtually any software developer.

We have long said that by enabling developers to use our services at a deeper level we would enable new kinds of services that we  could not develop ourselves. Now we are seeing exactly that. Kudos to Terry Reese for building new capabilities into an already stellar application.

I’m pleased to report that we have an article in the most recent Code4Lib Journal! The article, Thresholds for Discovery: EAD Tag Analysis in ArchiveGrid, and Implications for Discovery Systems is based on an analysis of how EAD (Encoded Archival Description) is used in the ArchiveGrid corpus. We go beyond that to look at EAD-in-use through the lens of discovery — how well is EAD currently meeting our objectives making our finding aids not only more discoverable, but more functional in discovery environments? In many ways, this article is a reaction to the many questions we receive about ArchiveGrid and why we do not provide a variety of indexes or advanced search features — the encoding and the data simply do not, at this point, support this functionality. We hope this paper can serve as the beginning of a discussion about some focussed efforts to improve the current situation.

With more than 120,000 finding aids, we believe this is the largest analysis of EAD done to date, and I’ll immodestly propose that the article is worthy of your attention!

Many thanks to Kathy Wisser and Jackie Dean for sharing their work with us — their earlier study of EAD usage provided an excellent model, and I’m grateful that they were willing to share their early results with us. Their article will be published in the next American Archivist, providing yet another look at EAD. Thanks also to my co authors: Bruce Washburn; the extraordinary developer behind ArchiveGrid and many other wonderful things; and Marc Bron, a PhD candidate in the computer science department at University of Amsterdam, who worked with us as an intern in the spring and did the analysis quite handily along with a number of other projects.

We look forward to your comments on the article, as well as your ideas for how to move forward to better thresholds for discovery.