I’ve noticed both an increasing onslaught of articles and studies written about society’s relationship with their mobile devices, as well as a personal increase in my proclivity to pull out my phone for short-term stimuli in those 30 seconds throughout the day in which I’m doing nothing. Examples of this for me are usually mindless scrolling through twitter, reading (only) a headline about how bad the other political party is, or the worse one, unconsciously opening an app that I had just closed five seconds earlier. So, being a curious data nerd, I decided to start tracking my mobile usage a few months back so that I can better understand my reliance on technology. What follows is the exploration into that data, with a specific focus on testing the hypothesis that:
too much of my mobile time is spent in short (5–15 second) chunks of time (which, I believe, is an indication of technological engagement that adds little value to my life)
Before diving into the data to check my hypothesis, I should point out that embedded within the above hypothesis are two underlying assumptions:
(a) Not that long sessions are inherently good (often quite the opposite for people), but simply that these very short sessions are probably bad. Plus, I already know through my data that my longer sessions are reading and that I very rarely use social media (outside of twitter), so I’m inherently assuming (for the time being) that longer sessions are an indication of either neutral or good use of technology (though, a future analysis will certainly look at the quality of these longer sessions, and I have already made a concerted effort to read more physical books this year instead of digital text). This is personal to my behavior and doesn’t necessarily apply to others.
(b) More broadly speaking, that we have an over-reliance on technology that’s harming us, as it:
- encourages a poor feedback loop of dopamine hits and short-term stimuli which leads us to constantly need to be entertained, even if said entertainment is short-lived and provides little value to our lives
- crowds out our ability to be present in the moment
- harms our ability to focus over long-periods of time and be creative
- gives us a false sense of reassurance that we’re “in the know” and involved in what’s important in life
- often results in bad physical posture
I would like to point out that I’m not suggesting we never use technology, but simply to use technology wisely and within our control, and not hesitate to pick your head up and be present in the world around you.
To set some context, here’s the high-level summary of my phone usage (as provided by Moment directly in the app). Also noteworthy context, is that Moment users average 52 pickups across 3 hours and 42 minutes every day, which may sound astonishing, until you get a broader picture of the public, which apparently averages 85 pickups across 5 hours per day. To put that last one into perspective, the average person is on their phone 31% of the time they’re awake, and equating to nearly 13 years of your life!
The data included in the below analysis is exclusively for 2017 (through mid-April), and includes 5,000+ “picked up my phone” logs, along with ~350 “daily app usage summary” logs. For those adept at the data side of life, I basically imported Moments JSON file into a postgresDB (I have a personal db that centralizes various data points), and then used an api to feed that data into a Jupyter Notebook.
To start, you’ll notice that of the roughly 95 minutes a day that I spend on the phone, approximately 35–40 minutes is dedicated to reading (Instapaper) — which is a habit I’ve been cultivating (nearly daily) for the past 7.5 years — and is a part of my morning calmness ritual of coffee, meditation and reading.
To move onto the data helping me test my hypothesis, it largely consists of logs (similar to a row in an excel sheet) for every time I picked up my phone to do something, and looks a bit like this:
Not surprisingly, it appears that a lot of my mobile pickups last less than 120 seconds. Though, these graphs don’t do a great job of telling a story, so I decided to bucket the logs into timed groups and see where they landed:
In this case, 25% of the time spent on my phone lasts less than 15 seconds, and while some of that is probably valuable time (e.g. quickly responding to a text), it appears my hypothesis is correct, or more accurately, I have not been able to disprove my hypothesis (from a somewhat loose scientific perspective). If we dive down a bit deeper, we’ll notice that ~7% of the time I pickup my phone, it lasts less than 5 seconds. So, somewhere between 3 and 11 times per day (out of 44), I engage with my phone in a way that likely lacks any real meaning (e.g. < 15 seconds). Sounds like an opportunity to gain back some of my day.
The final item that I wanted to review is the variance in short-engagement pickups, and it appears there is indeed a high variance:
And also the 7-day trailing average for <15 seconds vs 90+ seconds (not too surprising considering it’s a percentage and therefore constrained to 100% if you also include the time between 15 and 90 seconds):
While I usually like to pair action with any analysis, in this case I’ve only scratched the surface and will have to simply keep in mind that somewhere between 7% and 42% of my pickups lack meaningful engagement. However, I find that self-awareness of your behavior is often enough of a driver to change behavior. Though, one specific way I have started to combat poor phone use is by leaving my phone in another room so as not to engage in dopamine-driven phone sessions. I’ve been toying with the idea of no technology Saturday/Sunday mornings, but I’ll have to see how the data plays out the rest of this year.
In the future, I’ll probably perform further analysis to better understand my technology use, likely in these areas:
- Attempt to pair which apps correlate/cause short and long-term phone usage (not possible with the current data set) — for instance, I do a lot of curious google searches driven by engagement in the real world
- Use the lat/long data to see if ‘home vs work vs out’ plays a large role in certain mobile uses
- Combine with other data, such as my daily habits, to see if there are correlations between phone use and achieving habits, or if there are correlations between mobile usage and productive work sessions (such as long, continued software development time, as measured by RescueTime)
- Develop a machine learning model to predict which times that I pick up my phone may lead to productive vs unproductive mobile sessions, and send a gentle reminder (e.g. push notification) to myself encouraging an alternative activity, like meditation or sending my fiancé a positive affirmation