Envisioning Social Applications of Library Linked Data

Editor's Notes

• #2: This talk discusses two streams of innovation on the Web--the Social Web and Linked Data--and explains how bringing them together can move library services to the 21st century. Long a mainstay of social networking sites, social features such as user-contributed content and crowd curation are becoming more important for libraries, since they help to build communities (which is one of the functions of libraries) as well as to provide users with better and more relevant services. Linked Data has been widely adopted by businesses and media organizations, as well as by large libraries, in order to facilitate data interoperability on the Web. Combining the two provides an opportunity to create new applications of library data. In particular, we are interested in applying linked data in order to create novel, social uses of library-related information--uses that might support collaboration or that depend on information contributed by a large number of people. The presentation will consist of a short introduction to Library Linked Data and social features related to library data, followed by an overview of use cases collected by the Social Uses cluster of the W3C Library Linked Data (W3C LLD) Incubator Group. The core of the presentation will provide an in-depth look at a few of these envisioned use cases: social annotation, peer-to-peer bookswapping and social recommendations. The goal is to create interest in combining new technologies and to start a discussion about how to bring these and similar use cases to fruition. The presentation is based on authors' work in the W3C LLD Incubator Group and the output of its Social Uses cluster in particular.
• #3: Edit the Real-time Google Doc ! - can collect information (or just make notes) related to the talk
• #4: At the start of the talk, give people a link to collab-edited document - where they can collect information (or just make notes) related to the talk - as a lighter, real-time equivalent of the wiki links we have at end of the talk
• #5: roles of linked data (how linked data can be used): - publish linked data - exchange data b/w systems - import linked data = enrich [existing data] with data from outside sources
• #6: May 2010 -- October 2011 W3C LLD XG Social Uses cluster http://bit.ly/lld-social-cluster Main report (benefits/current situation/obstacles): http://www.w3.org/2005/Incubator/lld/XGR-lld-20111025/ Use cases report: http://www.w3.org/2005/Incubator/lld/XGR-lld-usecase-20111025/ Vocabularies & datasets report: http://www.w3.org/2005/Incubator/lld/XGR-lld-vocabdataset-20111025/
• #8: http://www.w3.org/2005/Incubator/lld/wiki/UseCases#Cluster:_Social_and_New_Use In this talk: - I will show examples of some Social Applications for Books and Libraries - which are interesting by themselves - show how they can benefit from linked data See the collab-edited doc + wiki pages of individual use cases for more info
• #9: Socian Annotation use case http://www.w3.org/2005/Incubator/lld/wiki/Use_Case_Social_Annotation Marginalia image: http://www.flickr.com/photos/58558794@N07/6796113096/ l icense: CC-BY
• #10: Marginalia (on paper) List of tools: http://en.wikipedia.org/wiki/Web_annotation Shared: CommentPress,
• #11: A reader (1) reads a book, (2) highlights a sentence and/or (3) adds a note All this interaction created information The information can be published on the Web and to social networks (e.g., as linked data) Aggregate and use the info: - item-centered view - people-centered view
• #12: Annotations may refer to objects at different levels of granularity. Some annotations describe the work (e.g., a book) as a whole while others annotate chapters or short sections of the text (paragraphs, sentences, words, pages...). Annotations may quote the content they refer to and add some additional information about it (e.g., reader's comments). Bookmarks and linemarkers -- which mark position but do not add text -- are also considered annotations. --- Logical fragments
• #13: Linked data can be used for: identifying works to annotate fine-grained annotation -- pointing to fragments of a given work publishing and aggregating annotations 1. Many publications already have URI identifiers and some have metadata available as linked data. These can be used for: using common identifiers to point to works, facilitating annotation aggregation; bootstrapping annotation applications by filling in information from the available work metadata. 2. A way to define fine-grained locations within / fragments of works is needed. URI schemes or RDF properties developed for identifying locations / fragments of works can be created this is an open problem (see Problems and Limitations in the use case doc ) 3. By publishing annotations as linked data users make them available for easy aggregation and reuse
• #14: Peer-to-Peer Bookswapping use case - Connect people and books (both paper- and e-books) - Enable individuals to share e-books easily, and legally (i.e. according to license restrictions). - Raise awareness about existing lending options and their limits. http://www.flickr.com/photos/interlude22/5384934730/ - "You know you work in Shoreditch when Rework is on the book swap shelf of your local cafe" Book swapping: - friends - hostels, cafes http://www.w3.org/2005/Incubator/lld/wiki/Use_Case_Peer_to_Peer_Bookswapping http://www.bookswim.com/ - book lending a'la netflix http://bookcrossing.com http://www.librarything.com/work/9279041/get#swap People come to a location (e.g., a cafe) and can participate in a bookswapping "club". They take a book from the shelf (a library), leave a book instead and record a note in the journal. Downsides: - having to go to a location limits it - some people may not be ok to be bothered to go to some place - they may have a limited selection of books, not what is interesting to you
• #15: Web-based paper book swapping A more advanced version would be one which lets people swap books without the need for a location. Virtual communication portal would match people with books to people with the interest in reading these books: - similar to sites selling used books, but without the selling part to it -- more about exchanging books Benefits: - can find books that interest you (IF the service grows to be large enough or to have the books you are interested in) - not that limited by a single location - less need for serendipity/chance/happening across stuff - can connect people with similar interests Downsides: - needs some level of trust (?) - shipping costs/trouble of arranging a meeting Next: - how can linked data help here? - automatic matching in the catalog - restrict to trades (they have a book I want/I have a book they want) - lots of different sites, don't really care which site you're getting from Role of linked data: - enrich the information about books (bring in info from other sources) - connect multiple book swapping sites - or desktop software to book swapping sites (people can define offline what they are looking for and then bring that data to the site via linked data ==== - Book Crossing - a few of the other sites listed here? http://www.librarything.com/work/9279041/get#swap
• #16: Commercial services: http://ebookfling.com/ http://www.booklending.com/ Match Kindle borrowers & lenders (also Nook ones) Besides - e-lending from libraries More e-books available: - free sample books from publishers - public domain books - Creative Commons licensed books - more CC-books being unglued (unglue.it)
• #17: Downsides of RL swapping: - having to go to a location limits it - some people may not be ok to be bothered to go to some place - they may have a limited selection of books, not what is interesting to you Downsides of web-based RL book-swapping:: - needs some level of trust (?) - shipping costs/trouble of arranging a meeting Linked Data may not be able to address the shipping cost / lending restrictions problems - but it can help address the information problem
• #18: distributed lists of who has what + wishlists - collab between sites - ...
• #20: Goal: Provide users with recommendations or search rankings based on information available about item popularity and user activity. Krebs says: " I would like to judge an item by the company it keeps -- its network neighborhood. Who points to it? What communities is it a member of? Is it central in the community? Does it bridge communities? Are there equivalent alternatives? " networks drive recommendations understanding what people like, buy, ... helps find other things that are similar 2 types of networks that are interesting: - people networks -- network of *things* derived from customer interactions - networks of things -- derived solely from themselves (Statistically Improbable Phrases, bag of words, "equivalent books" --...) -- http://orgnet.com/booknet.html -- " structural equivalence" -- two books that could substitute for one another photo: Valdis Krebs - http://orgnet.com/divided.html See also: http://www.edfinn.net/2011/11/emory-discourses.html
• #21: librarything or goodreads http://orgnet.com/booknet.html http://orgnet.com/divided.html
• #22: Clickstream analysis of scientific journal usage from library use logs (c.f. the keynote of day 1, ELAG-2012) - understanding the connections that readers make http://www.w3.org/2005/Incubator/lld/wiki/Use_Case_Social_Recommendations#Use_Case_Scenario --- A ) Generic case aggregate-level data B) Individual personal- / individual-level data see Use Case wiki for types of social data
• #23: Volume anonymity sharing & reuse is important -- Examples: University of Minnesota http://journal.code4lib.org/articles/501 Oregon State -- Summon catalog Image: http://journal.code4lib.org/media/issue14/katz/image02.png http://journal.code4lib.org/articles/3655 --- JS: individual libraries don't have Amazon's volume. Also want to anonymize data -- so sharing is key. That makes transporting data important. Also makes public data sets (like circ transactions) useful. I think that *having* the data is more important than having LD. But the LD makes it more usable. Especially since different ILS's have different internal formats -- would need to map between those. --- How your library could do this? Making it even better with linked data: - share the data with others (= mainly aggregate data) - bring in more data to show richer results, recommendations
• #24: *what kind of public data sets are there/could we envision? ** linking between the items, interactions on those items, social networks around those items Q: how is it different from just a simple Amazon-type recommendations use case? JS: individual libraries don't have Amazon's volume. Also want to anonymize data -- so sharing is key. That makes transporting data important. Also makes public data sets (like circ transactions) useful . SPARQL: hidden data but queries over the data
• #25: Named entity extraction Recommendations: - co-appearance in books? - nets of books, things, ... Can open up this data - do recommend / point to other books mentioning same things - can do things-network analysis - ..
• #26: * “ small linked demons ” :-) *
• #27: c.f. Valdis Krebs research on Amazon book networks - http://www.thenetworkthinkers.com/2012/04/next-big-thing.html (!!!) - earlier work: http://www.orgnet.com/divided.html (re book purchasing patterns) - that's something you can do with just the information that's been available on Amazon website for years - cue-in for: imaging what you could do with more, richer info from Small Demons and other cool services what's the value here / things you could do if linked data were added to Small Demons? - when you set data free people do stuff with it (!) - "here's the things I could do with this data and you can do this too" - essential value proposition of open data, linked data: reuse/sharing means anybody can make anything -- find new connections to things - "social networks" of *anything* - making the work about the decisions rather than the data management (LOD store AND the right kind of metadata --> questions you ask, interfaces you can put on top; "data" is already there, "metadata" is already there) c.f. Tim Berners-Lee's TED talk - http://www.ted.com/talks/tim_berners_lee_the_year_open_data_went_worldwide.html - "here's the cool stuff people do with [linked] open data"
• #28: "Raw data now!" http://www.ted.com/talks/tim_berners_lee_on_the_next_web.html
• #30: * empower / facilitate the community - to do stuff with the data! * building a community to move the initiative further: - collab-edited doc: -- put the links there ourselves -- encourage people to add links to the wiki page - communication channels - W3C Community Group? - hashtag for social uses of library linked data