Background
what makes a good use case?
- which of them is really linked data enabled, vs. what you could do with MARC if put in a big database
- but it may also be valid to show what can be done with linked data, even if it could be done without
- examples tying library data together with faculty profile information, archival information, and other sources not described in MARC
- and working across institutions
- enables a target audience to "get it" – LD experts, librarians, university administrators, scholars, the mainstream media
what intersections of our data sources will be strong enough to support compelling use cases?
- we have good bibliographic data
- we have usage data
- we have information about our faculty (their publications, but also potentially grants, research groups, patents, facilities they use, other research resources, datasets they have produced)
- some amount of organizational and classification data – what's shown up on the reading list for a course, or been included in reference consultations, or has been identified as a classic text in a research guide
- BUT there are companies trying to sell us information they have indexed, gleaned, and sometimes disambiguated
Use Cases
| As a... | I want to... | Satisfied when... | Comment | Importance (1-5 5=highest) | |
|---|---|---|---|---|---|
| 1 | As a researcher exploring a new field, or as a reference librarian | I want to find what is being used (read, annotated, bought by libraries, etc.) by the scholarly communities not only at my institution but at others, and especially to find sources used elsewhere but not by my community | I'll be satisfied when the result of a search for a subject or for a particular work makes suggestions for further exploration that are both relevant and surprising | Demonstrates scholarly communities learning from one another across institutions | 5 dw
|
| 2 | As a faculty member preparing the readings for a new course (or refreshing an existing course) | I want to see what works are being used by colleagues in different institutions, by searching for a course subject, or by suggestions based upon a work that I know about. I would like to be able to browse to the other works being used in a particular course, and also to see the works that are on the periphery of the standard works within a subject. | I'll be satisfied when a tool enables me to do this. | ||
| 3 | As a faculty member or librarian | I want to create a virtual collection using online materials in multiple collections across multiple universities. I need a tool that will let me browse fluid, discovering items that otherwise would have escaped my attention, and easily creating an online exhibition. Information about user interaction with this exhibition – unique visitors, but also which items they click on, in what order, how long they spend with each – should be fed back into the universities' systems to help inform future people browsing the collections. | I will be satisfied when this tool exists, and usage data can be usefully collected from the exhibition. | ||
| 4 | As a researcher | I want to identify colleagues at my school who have taught at or were graduated from a partner school, and see what work they have done on the topic they pursued at my school after they left my school. | Why would I want this? dw | ||
| 5 | As a student | I want to browse a geographic map and see annotations automatically added that show me relevant information (library items, archival items). I want to know that new information is added in close to real time. | I will be satisfied when that works | Maps are a gnarly type. And how do we get the geolocated items? And this isn't very fresh. Is there a way to show relations among the pins? Or something? - dw | |
| 6 | As a university dean | I want to works created by my faculty to be highlighted in the OPAC. This includes works by any author who has ever worked at my college. | I will be satisfied when that information is accessible by the OPAC, although it'd be nice if the OPAC actually used that information. | ||
| 7 | As a media person looking for a local expert | I would like to see which subject areas the faculty of a university are focused on, by seeing the subjects on which they are publishing and teaching. | |||
| 8 | As a librarian | I would like to assess the strength of my university library's collection by seeing for every LC subject class how many resources we have, how heavily they're utilized, and how much research is being done by our faculty in those areas. This could highlight potential areas of the collection to enhance or winnow. | I will be satisfied when I have an analytics tool that lets me make such inquiries. | ||
| 9 | As a researcher working on a particular academic figure | I would like to be to see the works written by the students of a particular figure. E.g., what have the people who took actual classes with Buckminster Fuller written? Then I'd like to see the citations for those students' works clustered by subject area. | This requires data we may not have. - dw | ||
| 10 | As a researcher working on a particular academic figure | When I search on Buckminster Fuller, I would like results ranked not by relevance but by "community relevance," i.e., by how often those works are used by my university's communities. | I will be satisfied with Stacklife | Doesn't demonstrate linked data - dw | |
| 11 | As a teacher and as a librarian | I would like to easily create a "shelf" (a list, really) of supplemental readings for my students, and have the relationship among those items – namely, that they all have some reference to the main sources – be fed back into the Library's data set so that others can benefit from my intellectual work of clustering them. I would like the provenance of that clustering maintained. This information could then be presented through a browsable discovery system. | |||
| 12 | As a researcher | I would like an alternative to Amazon's "people who read this also read that." The alternative should be focused on finding results relevant to me as a researcher, rather than works a vendor is trying to sell me. In fact, I would like to see side by side lists of Amazon's clusters and my Library's clusters so I can see just how much the Library's kicks Amazon's skinny butt. | 5 dw | ||
| 13 | As a student in a MOOC | I would like a tool that lets me create a "shelf" of works as I do my research. I'd like to be able to see other student's shelves for a particular assignment, and I'd like to see an aggregated shelf that shows me the most shelved works. I'd like clicking on the items in these shelves to reveal other student shelves on which they appear. I'd also like to be able to see each work in a shelf other books related by subject, | Could be SQL - dw | ||
| 14 | As a researcher and teacher | I would like to be notified when there are archival items that might bear on a topic I'm researching or teaching. I would like this to work at the component level, not only at the "box" level. | This would require devolving metadata from the box to the component level; Harvard has done work on this and has a set of EAD components - dw | ||
| 15 | As a librarian | I would like to enter any form of an author's name and have authority control | (I don't understand this - dw) | ||
| 16 | As a librarian | I would like help building my collection by seeing what is being used by students and faculty, and what's being used at other universities | |||
| 17 | As a researcher or student | I would like to be able to ask questions about the relationships among library data using something like natural language (although I'd be happy to accept some common sense restrictions) and get back interesting results. For example, "What books were written by Cornell biologists on topics that Harvard and Stanford biologists rarely write books about?" "What books are available about communication technology that don't use 'Communication' in their title or subtitle?" "Create a timeline of science books that charts them by how many of those books have illustrations." "Find all the anthologies that have chapters by both Marshall McLuhan and Neil Postman." "Create a table that shows how many genetics books were published in England per year versus how many were published in the United States." "Find me books about Christian Fundamentalism read (or assigned) in Divinity schools in the 1990s but less so after 2000." "Show me chronologically the usage in medical schools of books under the headings of both vaccination and autism." "Map by publisher location the clustering of books about American slavery since 1640." (Library Graph) | I will be satisfied when there is at least a limited vocabulary and syntax that I can master and get back interesting results. | (We could limit the NLP issue by creating a pick list of operations, relations, and outcomes. Or maybe there's an Open Source NLP library we could use. I dunno. - dw) | 5 dw |
| 18 | As a researcher or librarian |
| (Explanation needed -dw) | ||
| As a librarian working on acquisitions | I would like to find additional materials that build on our current holdings – additional works by the same authors or about the same topics. | I will be satisfied when I have a tool that lets me explore these ways | |||
| As a researcher | I would like to be able to build virtual shelves of disaggregated materials. I'd like to be able to use them in classes, and to make them public to provide context to what I'm posting. | ||||
| 19 | As a researcher or student | I would like intelligent term expansion:
| |||
And a slightly more narrative approach is okay.
see examples of both short and narrative form (pertaining to LibraryCloud, not linked data)
Characterize and prioritize
- order and sequence
- essential vs. edge case (importance/priority)
Examples
Identifying local authors
- When searching a topic in the catalog, identify which publications are by a home institution author
- Scholarly or promotional?
- scholarly – I can search the catalog and find a local author who has written on that topic
- promotional – this is someone who's a Stanford author or who graduated from Stanford
- potentially lots of flavors – have they ever been a professor at Stanford
browse by institution and show where their faculty are focused
- bringing in the subject classification
- Katy Börner map of science visualizations
- what would the use case be --
- being able to assess the collection strengths by subject (including adding classifications to some source data that do not currently have it)
- being able to compare subject coverage and strength to usage of library resources – the stack score, showing what is ignoring
- self-referential use of books reinforcing the head and the tail
- self-citation within the university
- over-reliance on top lists
- I'm trying to put together a reading list and want to see what else is being taught (when don't have access to syllabus data)
- why triples?
- opens it up to many forms of analysis and/or visualization
- being able to use the same approach and format across universities
archival use cases
- searching the discovery interface and finding primary source data applicable for teaching a class that don't currently know exists
- use the description and/or reading list of a course to find related archival materials
- stronger linkage between Stanford authors writing in Dutch history and their research descriptions in the 18th century and digital collections or archives
- doing better than folksonomies and plain text strings
- cross referencing what holdings are, what's circulating, what faculty are publishing, what's taught in courses
- to affect acquisitions decisions
generational impact factor – what have the students of a person working in this domain published?
library workflow use cases
- enter any form of an author's name and have authority control
- CuLLR broadening – virtual library or unit/subject library use case – curatorial
- identify the shelf of engineering reference handbooks
- tagging for a reading list (not classification)
- acquisitions support – helping selectors find additional materials building on current holdings
- additional works by the same authors, about the same topics in the same time period or place
- building virtual collections based on disaggregated materials
search inference
- aliasing of terms through language
- term expansion that is more entity-based
- using linked data to power "see also" panes
- examples where string comparison is not enough (e.g., jaguar)
- weeding out false matches
- end user tool
find books about Islam read in divinity schools in the 1990s but not now
all the books used in medical schools under the headings of vaccination and autism
work-based discovery
breaking apart composite works rather than the bibliographic entity
TOC of a FestSchrift
Musical
I want to find a chart to "something's gotta give" that wasn't
Beethoven's 6th bagatelle
find the Youtube performances of that bagatelle
hitting IMDB or Amazon to make our media collections better used
other movies that Paul Walker was in
finding the most popular thing for wildly underspecified searches
bringing in stackscore
authority-driven resolution
- example?
- what can you integrate that you can't integrate now
usage data
- library usage data is too sparse – aggregate from multiple libraries
- interesting differences among institutional usage data
- consider these 4 works which are read heavily at Cornell but not in the top 20 at Harvard
- citation count vs. usage count (is there enough at the monograph level? no, but can't really get into article-level usage, just at the journal level)
- a report of subject strength compared to usage by subject
- what are the requirements for making circulation data more useful while respecting privacy – can we do web-scale analysis on the amount we do have
- e.g., religious studies, where aggregating data makes it less risky
crowd-sourcing
- solicit feedback (as on Google graph) from users where the linkages suggested are not appropriate (or are the sources uncertain enough that researchers will reject the whole effort out of hand)
- Google did it with Freebase to identify entities
a graph search engine based on the Facebook model
- all the books co-authored by David who have been cited by authors with expertise in discipline
- SNAC approach in archival materials – establishing step relationships between people in archival collections
- finding all the books by Buckminster Fuller read by divinity students
- finding the chain of publications going back in time through citations
- Tom - there are citation mapping services and we're poor at doing them because we don't have the data
- Lee Giles at Penn State as citation mapping data in CS
killer demo – I start a search with a known item (Origin of Species) and takes advantage of the linked data graph to brings back:
- a better relevance ranked set
- a more diverse set ranked across different entities
- the best books by Darwin or by colleagues (related to the author)
- related by subject or species
- related by place
- reviews
- being used in courses
- most used overall
very dynamic sources
- books most relevant to the trending topics in the NY Times
forgotify serves up spotify songs that have never been performed before – but they can only be performed once – forgotten books, or books that have never been checked out?
