4 ms·

After reading and heavily agreeing with this post, I came to the conclusion that either goodreads is not really trying, or -more likely- the data is just not go

After reading and heavily agreeing with this post, I came to the conclusion that either goodreads is not really trying, or -more likely- the data is just not good enough to make decent recommendations. There are so many biases in the review data that are impossible to fix in any kind of sparse matrix recommendation algorithm. For those who want to try anyway, it might be worth downloading an existing dataset (1) (104 million reviews) and try, before worrying about scraping and api limits. The only solution (in my experience) is to get some other way of quantifying content, like Spotify does by manually labelling tracks. After some ddg I found storygraph (2), which does this. Its search engine is quite impressive, might be worth trying. [1] https://sites.google.com/eng.ucsd.edu/ucsdbookgraph/home https://sites.google.com/eng.ucsd.edu/ucsdbookgraph/home [2] https://beta.thestorygraph.com https://beta.thestorygraph.com

kobe_bryant 6y ago
the problem is that when you step out of genre fiction, you cant just recommend a book with a similar plot or setting, it needs to have a similar style of writing and sensibility and thats very hard to determine. the best way is friending/following people and learning their tastes compared to yours e: if you go to a specific book and look up similar books, goodreads actually does a pretty good job https://www.goodreads.com/book/similar/1994351-j-r https://www.goodreads.com/book/similar/1994351-j-r
- lstamour 6y ago
  I would think that Amazon with Look Inside The Book and Kindle eBooks has enough data to do this, if anyone can?
- sfjailbird 6y ago
  > the best way is friending/following people and learning their tastes compared to yours If only goodreads could somehow find others who like the same things you do and use it for recommendations! Joking aside, they might already be doing that, but if so, they suck at it. They have enormous datasets of people and their tastes, yet their recommendations always seem to be simply matching genres that you happened to like a book from. You liked Watership Down? Here are 10 other books with talking animals! If you and I can successfully identify people we share tastes with, then goodreads should be able to, too, they have even more data to base this on.
prepend 6y ago
Thanks for posting this. I googled around and came up empty. While I wish companies did more open data posts, I’m glad that researchers are filling the gap, sort of, by making data sets and analyses like yours.

kobe_bryant 6y ago

the problem is that when you step out of genre fiction, you cant just recommend a book with a similar plot or setting, it needs to have a similar style of writing and sensibility and thats very hard to determine. the best way is friending/following people and learning their tastes compared to yours e: if you go to a specific book and look up similar books, goodreads actually does a pretty good job https://www.goodreads.com/book/similar/1994351-j-r https://www.goodreads.com/book/similar/1994351-j-r

lstamour 6y ago
I would think that Amazon with Look Inside The Book and Kindle eBooks has enough data to do this, if anyone can?
sfjailbird 6y ago
> the best way is friending/following people and learning their tastes compared to yours If only goodreads could somehow find others who like the same things you do and use it for recommendations! Joking aside, they might already be doing that, but if so, they suck at it. They have enormous datasets of people and their tastes, yet their recommendations always seem to be simply matching genres that you happened to like a book from. You liked Watership Down? Here are 10 other books with talking animals! If you and I can successfully identify people we share tastes with, then goodreads should be able to, too, they have even more data to base this on.

prepend 6y ago

Thanks for posting this. I googled around and came up empty. While I wish companies did more open data posts, I’m glad that researchers are filling the gap, sort of, by making data sets and analyses like yours.