Personal analytics

I was impressed by a recent post by Stephen Wolfram discussing his “personal analytics”; that is, data that is automatically collected about ones self and then analyzed for interesting trends. Wolfram has been capturing such data for two decades and some of what he had to report was fascinating, especially if you like numbers. One of the plots Wolfram presented was a timeplot of the email he’s sent and received. That was data that I had readily available, at least for my day job, so I constructed a similar plot for all of the work-related email I’ve sent since January 1, 2000; that is, for the last dozen years. Keep in mind this is just the mail that I’ve sent, not mail that I’ve received. In that 12 year period, I’ve sent over 47,000 email messages. And in looking at the plot, you can see some pretty cool trends:

Email Analytics.PNG
1 dot for each email. Click to enlarge


The times plotted along the y-axis are in eastern time. So the first interesting thing that this plot shows is my move from Santa Monica to the Washington, D.C. metro area back in the summer of 2002. If you look at that time frame, you’ll see that I start sending emails a few hours earlier in the day–because I’ve changed time zones. I also start going into the office later, around 7am, whereas I was getting into the office between 5:30 and 6:00am in Santa Monica. From 2003 to about summer of 2008 I was pretty consistent in the times I started work. Then in the summer of 2008, it starts to get closer to 8am. Since then, it has gradually moved back toward 7am, although you can see scattered throughout the dozen years times in which I was sending email in the middle of the night.

From 2000 to 2002, there is a kind of white band at the 3-4pm range–my lunch hour in Santa Monica; and this band shifts to the noon-1 pm range when I moved to Washington. This plot demonstrates that I’m fairly good about keeping my lunch hour to myself. It’s when I get reading done and it’s a good mental break during the day.

Scattered throughout the plot are a number of “vertical” stripes. Many of these represent vacations, when I was away from the office and not sending email. For instance, there is a band in the summer of 2007, when I was in Europe for nearly a month. There is another one in the morning hours of the summer of 2009–when the Little Man was born and I was taking the mornings off for a while. And another in the summer of 2011, when the Little Miss was born and I took a month off.

I find plots like these fascinating. Stephen Wolfram has taken this kind of analysis much farther than I possibly could. That said, I capture quite a bit of data. Now that it has come to my attention that there is some useful information that can be extracted from these type of personal analytics in the aggregate, I’ll report on more of my own as I have time to look at it and analyze it.


  1. I feel obliged to ask if you used Mathematica for this. (Looks like Excel.)

    My email data would be boring. But for the past two years I’ve been keeping track of writing time, and have been slowly working on a program to generate nice reports about it. Someday that will be very interesting, at least to me.

  2. Elizabeth, no, while I own Mathematica, this was easy enough to do in Excel, once I figured out the quirk for plotting time against date in scatter plots.

    Writing times is on my list, but I’m looking for ways to automate the collection of the data. I’ve seen some applications that track how much time you spend in various applications, including specific documents. Since at least one of these includes Scrivener (my writing application) I am considering giving that a try.

    The key to this personal analytics stuff is to make the collection of data as automated as possible. You don’t want to add tasks to your day.

  3. I log in Bento, by the Filemaker people, and it takes only a minute or two a day. But automation would be great. Especially if it could count shuffling index cards on my whiteboard or scribbling on paper as well as working on the computer or iPad.


This site uses Akismet to reduce spam. Learn how your comment data is processed.