Fiction->Mysteries->Historical Mysteries is an example of genres, not subject headings. There is a subtle difference--the difference between "what is this type" and "what is this about". LCSH does have a very structured hierarchy, but it was not intended for the kind of shelf browsing you seem to be interested in. Different taxonomies have different purposes, and it is quite difficult to repurpose a taxonomy or translate between taxonomies intended for different purposes. It will never work as well as a dedicated taxonomy. That's one reason that genres never worked well in LCSH. LCGFT was created to remove genres from LCSH. I think it might be useful to go further and split LCGFT into separate taxonomies for forms and for genres.
BISAC was created for yet another purpose. It is a combination of broad subjects (where LCSH focuses on very specific subjects) and fiction genres (where LCGFT includes both fiction and non-fiction genres plus forms). It works well in a specific setting--a bookstore type environment for casual browsing of both fiction and non-fiction materials. I'm not sure it is possible to translate LCSH into BISAC, at least not very well. You could probably do all right for the non-fiction categories, but LCSH really doesn't have fiction genres.
Our library did create a taxonomy for our libguides and our database list. The taxonomy is basically a cut-down version of LCSH, focusing on the subjects of greatest interest to the university, falling generally along the lines of departments and degrees. It works well for our purposes. But it does not include fiction genres, because specific genres are not an important category of study for the university. (Certain programs do study fictional material, such as film studies, but the subject terms we use are adequate for these interests.) Our taxonomy is designed for a specific audience--the faculty and students of our university--and works well within that environment. If you are looking for a combined non-fiction subject plus fiction genre hierarchy, I think it will be rather difficult to translate or cut down LCSH the way we have.
I would suggest that, before you begin working, you make some decisions about the scope and nature of the project. What is the intended audience for this taxonomy? What is the purpose of the taxonomy? Will it include fiction genres? Non-fiction genres? Forms? Subjects? How broad will the categories be? How deeply layered will the hierarchy be? After answering these (and other) questions, you can do a better analysis of whether it will be possible to translate LCSH into what you seek.
[log in to unmask]
> -----Original Message-----
> From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of
> Mark Watkins
> Sent: Wednesday, April 13, 2016 8:05 AM
> To: [log in to unmask]
> Subject: [CODE4LIB] LCSH, Bisac, facets, hierarchy?
> I'm interested to use the LCSH data contained in the Harvard Open Metadata
> project to provide some hierarchical browsing (e.g. Fiction -> Mysteries ->
> Historical Mysteries on top of a book database.
> I'm a library sciences newbie, but it seems like LCSH doesn't really provide a
> formal hierarchy of genre/topic, just a giant controlled vocabulary. Bisac
> seems to provide the "expected" hierarchy.
> Is anyone aware of any approaches (or better yet code!) that translates lcsh
> to something like BISAC categories (either BISAC specifically or some other
> hierarchy/ontology)? General web searching didn't find anything obvious.