Maps of Let the Great World Spin

5-instance rolling and cumulative mean centers of each POV of the novel.

What is going on here?

Let the Great World Spin is a 2009 novel by Colum McCann. The novel is split up into ten chapters, each with a distinct voice, or point of view, or “POV”. There are three additional, shorter sections. The first, called “Intro” above, is of the opening five pages of the novel. The other two, called “Walker,” focus on the tightrope walker, never named, in the novel. I have collapsed those two sections into one.

The automatically showing layer is the mean center of each POV. Clicking on a circle will reveal which of the 12 POVs is represented. The layer menu (top right corner) reveals the 5-instance rolling and cumulative mean centers, respectively, of each POV.

Clicking on the newly revealed moving mean centers shows you for which instance (and on which page) that calculation was added. This can be confusing in the rolling mean center, because it may suddenly jump southeastward, making it look like a very southeasterly place (Sydney, for example) was mentioned. However, the 5-instance rolling mean center calculates the mean center of the instance under observation, the two instances before, and the two instances after. A sudden jump towards Australia means that in two instances we will probably see one that is for a place in Australia. This can be confusing, but I think it is a fairer representation of a rolling mean.

Both sets of data show, then, how the geographical “center” of the novel moves over narrative time. For example, Ciaran’s POV shows a general movement from Dublin to NYC, stopping somewhere over the Atlantic, which fits the tone of his section of the novel (the first full chapter). But on a 5-instance rolling basis, we also see that Ciaran is all over the place, persistently thinking or mentioning places not in NYC well after moving there.

How was this made?

Geographical calculations done with the suprisingly flexible and friendly Turf.js. Currently disabled animated markers courtesy of Leaflet’s MovingMarker extension. Data is by Moacir P. de Sá Pereira of the NYU English Department and taken from NYWalker, a NewYorkScapes and #NYUDH joint.

The code for this page, short as it is, is available on GitHub.

This is part of an instance of “everyday criticism.” Experimentation is part of everyday criticism, and this was easier and more fun than yelling in python at ArcGIS.