A summary of this project is available here, via the knitted markdown document.
I looked at the Wikibooks data set on Kaggle and found that though there were a lot of books the data set was hard to do anything with because there wasn't much meta data.
I wanted to try to extract ISBNs from the data set for two reasons.
-
Having ISBNs can be used as a VERY roug proxy for the book having a certain lever of quality.
-
Having the ISBN will allow linking to other data sets in library catalogues, this means additional high quality meta data can be added to those books. Variables such as publisher, year published, author, subject etc. should be available.
I was able to extract ISBNs for books of all languages in the data set, including Japanese.
However the method only worked for a small percentage of books.