This page contains links to visualizations and data for the novel endings, which we modeled using 500 topics. I also modeling the full set of novels and the set of the top 30 novels using 500 topics, because I wanted to see if we could gather anything from using a consistent set of topics with different sets of data. Also included are new visualizations . . .
The topic spreadsheet for the endings. This is the spreadsheet which led us to notice that a lot of novel endings had a lot of topics 13, 324 and 393.
The most common topic words for the endings.
A new network visualization, replacing the old style, very wide graphs. This graph shows novels which are similar using our usual measure.
Try mousing over the circles and clicking on them. Also notice the box in the upper left-hand corner: try expanding it, them checking and unchecking the check boxes.
At some point in the future, I should be able to create a similarly interactive viewer for the output of multi-dimensional scaling . . .
A similar visualization showing novels which are similar using a relaxed definition of closeness (I used a relaxed definition here because our usual measure was so unsatisfying).
Another connecting novels to whichever novel they are closest to. I connect each novel only to whichever novel it is closest to, regardless of how close that is.
The upside-down, one-novel-at-a-time novel distance viewer.
The topic spreadsheet for all of the novels.
The most common topic words for all of the novels.
The new visualization for all novels which are similar using our usual measure.
The new visualization for all novels connect each to the novel it is most similar to.
The novel distance viewer for all novels.
The topic spreadsheet for the top 30 novels.
The most common topic words for the top 30 novels.
The new visualization for top 30 novels which are similar using our usual measure.
The new visualization for all novels connect each to the novel it is most similar to.
The novel distance viewer for the top 30 novels.
