Sifting Through My Mobile Phone Logs

Posted to Self-surveillance  |  Nathan Yau

When I was in NYC and my wife was in Buffalo, New York we talked on the phone almost every day, usually around ten in the evening. I was at my friend’s place one night, and at 10:05pm, my wife called.

The first thing she said was, “Where are you?”

I told her I was at my friend’s.

My wife quickly replied, “Ha! I knew it!”

Confused, I asked, “How did you know?”

“Because otherwise, you would have called me at exactly 9:58.”

Am I really that predictable? First it was the Chinese food, and now I had been accused of call time predictability. Of course there was only one way to put this dispute to rest — look at the data.

The Analysis

Luckily, my carrier, Verizon wireless, offers call logs in spreadsheet form. I was only interested in chats with my wife while I was in NYC, so I sorted all of my phone calls by time and got rid of the records that weren’t her. After some data cleaning and adjustments, I threw the data at R (a statistical computing language) with all of my might, and it kindly provided me the graph below.

Not Calling Her at the Same Time Every Night

The Findings

As expected, I didn’t call at 9:58 every night. In fact, the most calls were at 9:57. Ha. So there. Alright, maybe my call times were slightly predictable, but definitely not to the extent suggested. Most calls occurred some time between 9:30 and 10:30 with some scattered calls late at night and during the afternoon.

Data wins again. Data 2, over-generalization 0.

Have you looked at your call logs lately?


  • Lovely graphic – but a few minor criticisms:

    * what’s with the x limits?

    * the caption regarding the bin size could be a little clearer – the bins are 1 minute wide, correct? And it’s the time the call started?

    * some thing doesn’t feel right with the y-axis ticks – they look awfully like rotated bins

  • Yeah, I really should have extended the x-axis a bit to include midnight to 11:59pm instead of only my own call times. That’ll teach me to work on things late at night.

  • I really like these clever types of analysis…there is not enough appreciation of statistics as an *investigative* tool (using human traces etc.)! But give it a few years and we’ll see ;-)


10 Best Data Visualization Projects of 2015

These are my picks for the best of 2015. As usual, they could easily appear in a different order on a different day, and there are projects not on the list that were also excellent.

Pizza Place Geography

Most of the major pizza chains are within a 5-mile …

Unemployment in America, Mapped Over Time

Watch the regional changes across the country from 1990 to 2016.

Causes of Death

There are many ways to die. Cancer. Infection. Mental. External. This is how different groups of people died over the past 10 years, visualized by age.