Magic is Everywhere

finding patterns in the titles of fantasy books

Scroll down for information

Details
Link:

Magic is Everywhere

Created:

November 2016

Reason:

data sketches’ November month

Tools:

R, d3.js & Illustrator

Data:

Scraped Amazon for the top 100 fantasy authors. Next, retrieved each author’s top 10 books from Goodreads through their API

Description

In “Magic is Everywhere” the titles of ±900 Fantasy books have been investigated to look for trends in their titles. Each book’s title is represented by a circle and a line connects books by the same author. Areas of similar subjects in titles have been highlighted by colored oval areas and the corresponding term. The swooshes around each circle spell out the title itself (you can see an animation below the visualization on how to interpret these ).

Magic is the most common word in all titles. Followed by firedarkred & blood. In general, terms relating to unpleasant things are more common than those relating to positive aspects. Nevertheless, titles relating directly to magic can be found throughout the visual, indicating that the one thing that binds fantasy books together is, not surprisingly, Magic.

Creation

For the Books topic of November’s data sketches I wanted to look into Fantasy books. It’s really the only genre that I read (besides study books). I always feel that the titles of Fantasy books have a certain thing/rhythm in common. To get a decent set of fantasy book titles I started with the top 100 most popular Fantasy authors from Amazon. I had hoped to also get the list of ±10 most popular titles from Amazon, but that proved more difficult. So instead I used the Goodreads API, which gives a lot of information about any author, such the number of ratings per book.

After data cleaning I was left with almost 900 book titles. I used text mining and the t-SNE algorithm to cluster the books on general terms, such as battle, nature, blood or red. This gave each book a position in a 2D plane. Using a travelling salesman approach I connected all the books by the same author with the shortest line and that is really all that is visible in the final product. There is an interactive version in which you can hover and see who the author is (and their other books) and where I’ve annotated my favorite 3 authors. I’ve also made a print version that only keeps the most important aspects of the interactive, which you can see farther below.

Read more about the data, sketching and coding on the data sketches November write-up.