Visualized: Four Years On My Phone

Four years and 1.3 million records documenting my every tap and swipe. What could go wrong?

I've been recording my phone activity since June 2020. Every action, notification, and app tap - building a dataset of over 1.6 million records spanning 1,543 days. Four years later, I took a look at the data to unearth hidden patterns in my life.

Let's start with the million-dollar question:

How much time do I spend on my phone?

If I remove all records from the dataset where my device screen was off, I get a pretty accurate reflection of my screen-on time. The results are stunning:

A closer look at my daily usage over the course of all 225 weeks reveals a hidden trend:

Yep, that trendline is definitely going up. And although it's a bit exaggerated because of the non-zero vertical axis, there's an obvious increase in my screen time. Taking my daily average per year exposes this trend even further:

So what's driving this growth?

The dataset includes every second that I spent using any app on my device. This let me investigate whether any apps are playing an outsized role in driving my increased screen time. Here's a look at my top ten most-used apps. There's an outlier that emerges pretty quickly. Want to take any guess what app that is?

It's TikTok. The algorithm... It's just so good.

That chart is wild. Look how quickly TikTok cannibalizes screen time from other apps. TikTok's share of screen time relative to my top ten apps grew to 40% by August 2022. In that time, BBC News' share shrunk from as high as 9% to 3%, exposing how TikTok has become an key news source for me:

I'm using TikTok a lot. But when?

Here's a breakdown of the time I spend on TikTok throughout the day (in yellow) compared to my overall usage (gray). It exposes how my usage of TikTok starts tanking around 10pm when I get into bed, with my overal usage dropping around 11 as I hit the sack.

Activity starts in the morning between 7-8am. So it would seem that I get around 9 hours of sleep a night (spoiler: I don't - stand by for a different post on that soon).

That chart suggests I use my phone more in the evening, between 9 and 11pm. But that's not the case for all apps. Looking at the hourly distribution of different categories of apps reveals more nuance in my daily patterns:

For example, let's zoom into my sports streaming app usage throughout the day. Check out the spike between 8am and noon, followed by a smaller one at 6pm:

They reveal my obsession with the English Premier League, which airs in the mornings where I live, and the UEFA Champions League, whose games usually start around 1145am local. The smaller spike at around 6pm probably reflects me tuning into NFL games in the early evenings.

Contrast that with my use of communication apps (like WhatsApp and Gmail), which I use quite consistently from 8am-8pm:

So far, these hourly charts have obscured huge differences between the different app categories. This visual reminds me how much of my time is consumed by social media like LinkedIn, Instagram, TikTok, and Reddit. But what's most surprising is how relatively small my news consumption is. I figure that's because TikTok and Reddit are big news sources for me too.

I'd also expect that my news intake varies throughout the week. Which begs the question:

How does my phone usage change throughout the week?

Which apps do I use more or less on specific days? How do my news consumption, shopping, or food delivery habits evolve throughout the week? There's a lot to unpack here, and it's all about the weekends:

A category breakdown reveals what I'm up to on the weekends: ordering food, streaming TV and music, and snapping jpgs:

Could notifications be driving my increased phone use?

Ever eager to avoid any blame for my bad habit of increasing phone use, I began to wonder whether I'm doing it because I'm being urged to by my phone. Is my phone notifying me more often, prompting me to check and open my phone more often?

Before diving into the details, here's a snapshot:

As always, the details get more interesting. Here's a look at how my notifications have changed over time:

But that's not the full story. My dataset annoyingly contains no notification data between March 17 and December 12, 2023. Adding back in the data when it came back online on December 13 reveals that change might be afoot, with my daily notifications growing in recent months:

So what does this all mean?

I have no idea but surely nothing good, right?

A flawed analysis

This is

Under the hood

I captured the data using the paid version of the App Usage Android app and exported the information as CSV files which I analyzed using python (with the help of some gen AI tools). I polished some calculations in Excel and built the visuals using Datawrapper.

I generated the banner photo using Microsoft Copilot.