The Interactive Map of World Fiction

I am really interested in some of the attempts people have made to organize works of fiction into two dimensional mapped spaces. In November's #30DayMapChallenge I even created a very small Island of Fiction map demo myself to explore how fictional authors could be mapped onto the fictional territories of a fictional island. 

Most interactive maps of books that I have seen have used the idea of genres as the guiding taxonomic principle, determining where individual works of literature are placed on the fictional map. My own Island of Fiction map mapped fictional authors onto five genre states, 'Crime', 'Horror', 'Science-Fiction', 'Romance' and 'Fantasy'.

TheLibraryMap also uses the idea of fictional genres as its organizing principle. TheLibraryMap is an interactive map of over 100,000 books which are organized based on 'the genre and topics of each book'. On the map individual books are colored by 'genre and topics'. Unfortunately TheLibraryMap doesn't have a map key so you will need to guess which color represents which genre of literature. Individual books on the map are also sized differently - but again without a map key we are left to guess what the size of each circle represents (total sales maybe?).

TheLibraryMap provides no information on how each book was placed into its genre. Usually with such a large taxonomic interactive map entities are mapped using machine learning. But I have no idea if that is the case here. It would be really interesting to know more about how the books were organized on TheLibraryMap. It might explain for example how existential novels by Camus and Sartre end up sitting so near on the map to fantasy novels such as 'Throne of Glass' and 'Red Queen'. 

The HathiTrust Digital Map is a very similar interactive map which allows you to browse and explore the 14 million volumes in the HaithTrust's repository of digitized texts. This map not only provides a visual interface with which you can navigate the books in the HaithiTrust digital library it also includes a fascinating discussion about how the texts are organized on the map. A discussion which explores how organizing digital texts may require a whole new system of library classification.

The Library of Congress Classification system categorizes books into different broad subjects and then by sub-classes within each of these subjects. The HathiTrust Digital Map uses an entirely different method of classification. On this interactive maps texts are organized by the similarity in the vocabulary of individual texts.

The 'Read' section of the HathiTrust Digital Map takes you on a story map tour of some of the interesting patterns that have emerged by organizing this Digital Library by vocabulary similarity. The story map shows you how this classification system diverges or resembles subject based classification systems, such as the Library of Congress Classification system. It also explores some of the new 'clusters' of books that emerge when you classify by vocabulary similarity. New clusters of texts which have some syntactical similarity but which under a subject based classification system would be classified far apart.

This story map tour also provides a great illustration of how a digital map of a library can actually use a number of different library classification systems at the same time. On the HathiTrust Digital Map the texts are organized spatially by their similarity in vocabulary. However as you progress through the story map the texts are also organized by language and then by subject matter by applying different colors to the markers of books in different categories. In this way the map is able to pick out interesting clusters of texts which have similar vocabularies within subject classes, texts which have widely different vocabularies but are still in the same subject class or texts which have similar vocabularies but are in different subject classes.