The Patterns in the Dates

Getting dates has never been my strong suit, so I got a big list of them and started crunching numbers!

I started out with Excel, and quickly realized that it couldn’t understand the concept of a day before January 1, 1900. So I dumped Excel for Google Spreadsheets, and things went pretty well until I realized how poor Google’s charting abilities were. I got back together with Excel (with enough data that I didn’t need to worry about the dating problem), made some graphs, and, well, here we are.

One of those graphs has been made before by other people, so I’ll cover it first:

The dates in the corners of Only Revolutions’ pages don’t progress at a fixed rate. They don’t even match the season of the narrative! This chart tracks the distance between dates on consecutive pages. Sam’s side of the novel, for instance, starts out moving really quickly (about 300 days per page), but soon spikes and takes a dive down to about 60 days per page. Hailey’s narrative is almost the complete opposite: it starts out slow and takes a huge jump just before a small fall, resolving with a slow incline. I thought it was awfully weird how Sam’s line is much more sporadic than Hailey’s is.

Here’s something new and different:

This is graph of the time difference between the dates on either side of the narrative. Hailey and Sam start out exactly 100 years apart and steadily get closer (I don’t think the rate of change is ever more than one year at a time), until they even out at a stead y 41 years’ difference for most of the novel. The dates then eventually return to being 100 years apart by the end of the book. I’m not sure what event those little symmetrical bumps near the middle correspond to, but they’re pretty neat.

And now, onto some counting:

To continue my diabolical statistics binge, I calculated the months and weekdays of each date that appeared in each character’s narrative, and summed them up to look for something silly to contemplate as intentional. I’m disappointed to say that I didn’t find much — the most I can say is that the weekdays graph looks sort of symmetrical if you try not to focus on the screen. And that they both had the some number of Wednesdays. Another interesting result not indicated on any of the graphs is that the most popular dates in the overall chronology were November 22 and January 1, both appearing nine times in total.

This is the part where I tell you that there are no more pretty pictures. It’s time to briefly talk about about my process and presentation.

I derived many of my results from my original problem of looking for a pattern of advancement in the dates (i.e. the first graph). After that came a series of introspective “I wonder if there’s a pattern in…” questions, not unlike our collected enthusiasm over searching for prime numbers a few classes ago. Luckily, Excel only crashed five times, so my studies were relatively uninterrupted. I even almost forgot about going to class once.

I’ve tried to present my data while embracing qualities that happen to coincide with Tufte’s ideals. That is, I like to strive for pure, simple, easy-to-understand visualizations. My goal was to make my graphs both informative and aesthetically pleasing: an optimal combination, in my opinion, for drawing in readers enough to want to find out more.

WordPress’s botched presentation of my images (without clicking on them, at least) kind of detracts from that ideal. Oh well.