year-travel-work-home_SMALL

Every Minute of 2013

By / January 31, 2014 /

At A.R.O., we believe that context is the key to creating the next generation of awesome app experiences. In order to be more effective, apps need to know more about us, the users they serve. Saga, the essential lifelogging app, learns about users through their real behavior, with little manual input. Download Saga for Android or iOS to see how you stack up.

Every Minute of 2013

Saga started its beta program in 2012 and launched officially in May of 2013. Accordingly, we have an entire year’s worth of data for the first time — a record of where all of our users were for every minute of 2013 (when they had connectivity and Saga was enabled, about 97% of the time).

Click image to view details.

Click image to view details.

The above image shows every single minute of 2013 according to that data. Each row represents a single day, one pixel per minute. Times correspond to users’ local time zones, so that noon (the middle tick-mark at the bottom) is in the same position for everyone.

The image is color-coded according to user activity at each moment. With Saga, we have a number of ways of describing user activity, including home, work, travel, food, retail, and entertainment.

The first image is a summary of the three most common activities: blue represents home, green work, and red travel. For each minute, we count the number of Saga users involved in each activity and weight the respective colors accordingly. For example, pure blue indicates that most users were at home, green that most were at work, and teal that equal numbers were at home and work.

A few conclusions are readily apparent:

  • Most Saga users get to work at about 9:00 AM and leave around 5:00 PM, with travel before and after (the green bands tipped with red in the middle of most days).
  • Most people work only on weekdays; the regular horizontal bands represent the alternating weekends and weekdays of each week.
  • We can see obvious days off work (the wider bands of blue) at Memorial Day (May 27), Independence Day (July 4), Labor Day (September 2), and much of the latter half of December.
  • During the changeover from beta to our official release, we missed about a week of data in early May (the black band).
  • We introduced a bug just before Thanksgiving that over-counted travel (the red-tinted band in late November), and fixed it about three weeks later. While there was a great deal of travel around Thanksgiving, it was not enough to explain this three-week spike.
  • The large increase in number of users after we launched Saga in May resulted in less-noisy data; notice how much smoother the image is after the May no-data interval.
Click image to view details.

Click image to view details.

In this second visualization, we remove all home, work, and travel data and chart the next three most common activities: food (green), retail (blue), and entertainment (red). Again, a number of conclusions jump out at us:

  • Most people eat between 12:00 PM and 1:00 PM or between 6:00 PM and 8:00 PM (the vertical green bands at those times).
  • Eating times are less predictable on weekends; the green is more concentrated on weekdays.
  • Retail activity can happen any time between about 10:00 AM and 11:00 PM, with no strong peaks.
  • Both retail and entertainment activity increase on the weekends; the weekend bands are noticeably more purple.
  • Entertainment is less common than the other two and shows up only as a reddish tint in the evening and on weekends.

Mike Perkowitz is the CTO and Senior Data Scientist at A.R.O. Perkowitz is responsible for user modeling, designing A.R.O.’s scalable analysis platform, and managing research and innovation. He holds several patents, a Ph.D and an MS from the University of Washington, and an Sc.B from Brown University. Learn more about Perkowitz at http://www.linkedin.com/in/perkowitz.

We will publish more interesting data visualizations in the coming months. Add this blog to your RSS reader and follow us on Twitter to stay up-to-date.

Categories: Analytics, Data Visualization



2 Responses to “Every Minute of 2013”

  1. Jose says:

    Nice, however I would like to know if I can generate this graph (or similar) of my data.

    thanks

    • kittyireland says:

      We are planning on a few ways to let users visualize their own data. In the meantime, you can download your data as JSON or add it to your calendar.