This is more of a "That'd be interesting to have" than a "Please do this!". I'm not sure but it could be interesting to see how two novels correlate based on their readers i.e. something like "80% of those who read novel A are also reading novel B". If it were possible... we would be able to take our favorite novels and see which novels correlate the most with them! That is, we could sort novels by how likely they are to be similar to our favorites! Spoiler: Tl;dr I discuss how some of the calculations could be done. How it could be calculated? Many ideas come to mind; maybe using the correlation coefficient from stats (0), or a simple "readers of both A and B / readers of A" (1), "readers of both A and B / readers of B" (2), or a linear interpolation of (1) and (2) (weighted mean). If A is the novel we know and B the one we're looking for, then (1) would favor more generally popular novels, whereas (2) would favor more niche ones, I think. (0) Might be the most unbiased, but I'm not sure.. I'm not sure how computationally feasible would that be though : /
I have something planned but I haven't worked out all the details yet. I'll mark this thread as 'approved' and will come back to it. Thanks!