Department of Media Technology
BookSampo - Linked Data in the Service of Fiction Literature
Eetu Mäkelä
Department of Media Technology
Who am I? - Eetu Mäkelä ●
●
●
Researcher at Aalto University, Department of Media Technology, Semantic Computing Research Group Special interest in massive amounts of heterogeneous linked data and getting some use out of it Stuff I've done ○ ○ ○
http://www.kulttuurisampo.fi/ http://www.matkailusampo.fi/ http://www.kirjasampo.fi/
Department of Media Technology
Fulfilling Fiction Literature Needs is Knowledge-Intensive ●
●
Customer queries and needs relating to fiction are complex, necessitating broad knowledge of the content and context of fiction literature E.g. can you recommend books that are ○ ○ ○ ○ ○ ○
●
crime stories taking place in Helsinki? similar to the works of Kafka? typical examples of 1980s Finnish fiction literature? worse than their movie versions? prized in literary circles and deal with gender roles? poem collections dealing with despair and redemption?
What themes should I write about if I want to get government grants and literary awards? Department of Media Technology
The BookSampo Project ●
Collaborative project between Finnish public libraries and semantic web researchers to develop a shared database of Finnish fiction literature that is: ○ ○ ○
Curated on the web by volunteers in public libraries around Finland Capable of describing both the content and the complex context of fiction literature Capable of providing advanced automatic search and recommendation functionalities on the web based on the data entered
Department of Media Technology
Basing on Linked Open Data ●
● ●
RDF data model: a network of entities identified by globally unique identifiers -> Relates entities (authors, books, movies, publishers, awards, series, reviews, ...) to each other, manifesting the rich context of fiction literature -> Allows automatically merging in external information (e.g. from authority or movie databases). In BookSampo, geo-coordinates of places are imported, allowing map-based user interfaces Data model isn't pre-proscribed and can evolve Using ontologies, the computer e.g. knows that both doctors and nurses are medical professionals Department of Media Technology
Collaborative Editing Environment ●
The SAHA editor
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Collaborative editing of a rich, semantic network a major source of excitement ●
Allows explication of the silent knowledge of librarians into a common pool ○ ○
○ ○ ●
Newly added entities, along with their detailed information are immediately available for others to use Information about e.g. publishers need also not be repeated for new books, which adds an incentive to provide richer detail about secondary resources Any information added to an entity enriches also all linked resources, whether new or old Librarian recommends -field
Chat functionality for additional co-operation
Department of Media Technology
Librarian Experiences of the SAHA Collaborative Editing Environment ●
●
In the editor, a source of acclaim has been the semantic autocompletion functionality of SAHA Also valued has been the possibility for creating new keywords inside the project if no existing keyword suffices without this adversely affecting findability
Department of Media Technology
However, there have also been some problems ●
●
People not yet familiar with the network model and the expressive schema make mistakes: someone renamed the concept of 1960s to 1970s while intending to correct a book dealing with the 1960s to deal with the 1970s Sometimes, counterintuitive auxiliary resources are needed in the RDF model and the editor does not hide these, e.g. this book is number 17 in the Yellow library series
Department of Media Technology
Collaborative Editing Environment ● Booksampo has evolved a complex data model that contains for example: ○ Abstract works ○ Concrete works ○ Book covers ○ Authors ○ Series ○ Awards ○ Positions of trust ○ Fictional characters ○ ... -> expressing the rich content and context of fiction literature Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Department of Media Technology
Ontologizing the Fiction Literature Thesauri Kaunokki&Bella ● To offer intelligent services, the fiction literature thesauri Kaunokki (Finnish) and Bella (Swedish) were ontologized ● To be able to link to outside content, they were linked to the Finnish general ontology YSO ● The experience of the librarians who ontologized the thesauri was that this brought in a very welcome additional structuring to the vocabulary. Also valued was the fact that changing keywords to ontology references instantly turned the database from a single language to a multilanguage one ● The linking to a general ontology was also deemed extremely beneficial, as now, the management of general keywords could be centralized, while having the work still be immediately usable in the domain ontologies Department of Media Technology
Ontologies in BookSampo ● Other ontologies used: ○ KOKO ○ LINGVOJ (languages) ○ KOKO-Place ○ Ontology of times ○ Nationalities
Department of Media Technology
Department of Media Technology
Search allows for (besides simple search) ●
Quickly gaining an overview of the position of something in the whole field of literature ○
For example, searching for "Crime and Punishment" returns also all authors and works that claim inspiration from the work, all works that have been compared to Crime and Punishment in their reviews, all kindred works, adaptations and so on
Department of Media Technology
Search allows for (besides simple search) ●
Complex search using context and inference ○
○
For example, pretend that you sort of remember a book where there was a doctor character, there were inscriptions on the cover and it had won some sort of medal Now, typing "medal doctor inscriptions" into BookSampo matches the book Sinuhe the Egyptian, because its author Mika Waltari has been awarded the Pro Finlandia medal, the main character Sinuhe is a doctor and there are hieroglyphs on the cover of one edition of the book, which are known to be a type of inscription
Department of Media Technology
Department of Media Technology
Other End-User Portal Functionalities ●
Browsing functionality in the system allows for automatically locating other works interestingly related to the current work based on the rich metadata along with explanations ○
●
e.g. for the book Sinuhe the Egyptian, the system recommends Nefritite by André Chedid, with an explanation that both are historical novels dealing with the way of life in Egypt in the 13th century BC
Users may gather books into a virtual bookshelf, which may then be shared with others ○
This way, further silent information on book interrelations is pooled from users in a structured way
Department of Media Technology
Outside Use and Linking ● ●
CultureSampo The biggest Finnish news paper Helsigin Sanomat used the data in a cultural hack event
Department of Media Technology
Length of Finnish and Translated Detective Stories Joint project with Finnish public libraries on providing new services for fiction literature based on semantic context indexing ●
Able to answer questions such as what is the influence of Dostojevski on Finnish literature, what themes to write about if you want Finnish grants or awards and how the length of crime stories has grown internationally verus in Finland ●
Department of Media Technology
Number of Writers Publishing Only a Single Book in Their Career
Department of Media Technology
Department of Media Technology
Department of Media Technology
More questions answered based on the BookSampo data (and other linked data) ●
●
●
● ●
●
Do writers with different backgrounds (place of birth, gender, occupation) write about different themes? What themes should I write about if I want to receive government grants or literary awards? What are the most popular themes in fiction literature by year? Which Finnish authors have won the most awards? In fiction, which suburb of Helsinki is most strongly associated with crime? What would a most stereotypical Finnish novel look like?
Department of Media Technology
