A deep dive into your app usage

Google allows a way for you to easily download all your data on how you use your apps and phone (if you use an Android phone), and have a look yourself. While I have quite a lot of time at home at the minute, I thought it would be interesting to do so, and see which apps are the ones I use the most, and how my usage behaviour looks.

I decided to use R and RStudio to analyse my phone usage. R is an open source programming language that is specifically designed for data analysis and data science. I use it almost everyday, and it’s a fantastic way to perform data analysis. It has a huge range of free and open source packages for every type of analysis you might want to perform, and with its programming interface, there really isn’t anything you can’t do with R. You can perform statistical analysis, run Machine Learning algorithms, and even build entire web applications and dashboards.

We won’t be doing any of that today, but I will be using some of the most popular R packages to showcase R’s capabilities, as well as finding some interesting insights around how I use my phone!

To see a more detailed walkthrough of how R and RStudio works and how I created each plot, please have a read through the PDF at the bottom of this article.

Looking into my usage

Let’s start with something simple: a bar chart of my most used apps in 2020.

It’s not surprising to see that apps like WhatsApp, Facebook, Instagram and Gmail are high up the list. I often use LinkedIn and Relay for Reddit to stay up to date with interesting trends, news and articles. Plus, a special mention for 2048, basically the only game I have installed on my phone, that I’ve had for years and am still addicted to!

This is a good start, but let’s see if we can perform some more advanced analysis. Next I’ll look at how my usage changes from day-to-day during the week. I aggregate the data to day-of-week, and then compare total apps used each day to the number of unique apps used each day.

This is interesting because I can see that, while I use apps more at the weekend, I use a wider variety of apps at the start of the week – you can see that the red bars for Saturday and Sunday stand out, whereas Monday and Wednesday have a higher number of distinct apps used than any other days.

This is a good overall view of my usage, but I wonder if my usage has changed over the time I’ve been using Android phones.

While we’re not that far into 2020 yet, we can still see the spread is much broader this year, compared to 2019, when messaging on WhatsApp was a long way ahead. I’m also checking my emails more, it seems.

Now I’ll look at a slightly more sophisticated analysis. I wonder if my phone usage has remained the same over time, and at different times of the year. To do this I’ll count the unique apps used every week from the start of 2018, and then compare that weekly figure to the overall average in that time, as well as a seven day rolling average (which should smooth out any weird weeks where I use a large or small number of apps, compared to the average).

So despite some strange looking spikes (I don’t know what was going on Christmas 2018!) it has remained quite steady over the past couple of years. However, the red seven day average line has seen a slight increase into 2020. The shading indicates where I’m using my phone above the smoothed out average line, and the grey horizontal line is the overall average number of apps I use weekly (56).

Those large peaks and troughs are interesting to me. This makes me think that my variety of usage might be quite wide. I’ll plot a histogram and look at the inter-quartile range to see if I’m someone who’s behaviour does vary greatly, or if I’m quite consistent.

As it turns out, the familiar bell curve suggests that I’m pretty consistent with my phone usage. The peak of the distribution is very close to my daily average (nearly 26 apps used a day) and the light grey lines show the upper and lower quartiles are quite close around my daily average. I do have some outliers which are surprising – I can’t think of many days I use just one app in a day – maybe it was New Year’s Day! And that one day I nearly used 120 apps – who knows what happened then.

This leads me to want to look at times in the year my usage does (or doesn’t) vary. Now I’ll create a boxplot that looks at the month of the year, and shows the range of usage within those months. It should show outliers, so maybe I’ll be able to see where those low and high points are in the year.

As you can see, the relatively constant average use is shown here, with each median line (the middle of each box) are all pretty similar. However, we can indeed see that my least active days fall in October (the month of my birthday) and December (Christmas and NYE) – so probably not that surprising!

I hope you’ve enjoyed this look into my phone usage, and if you’ve made it this far, please take a look at the below PDF where I go into the detail of how I coded these charts in R – LINK, with instructions for you to try it yourself.

Open PDF.

Written by:

Jordan Peck Senior Data Analyst

Category:

What we think

Date:

30/04/2020

You may also like

What we think

  

/  24 Jun 2020

Is Safari blocking Google Analytics?

The short answer, is no. Eyebrows were raised this week when an Apple Insider and a Search Engine Journal post went live bluntly saying "Safari now blocks Google Analytics on sites". The reason for this assumption was a recent Apple press r

Read more

What we think

  

/  06 May 2020

Why now is the perfect time to spring clean your brand

This week I bought two window squeegees. I can’t say cleaning windows was the top of my priorities a few weeks ago, yet after spending quite a lot of time inside the house recently, that filmy layer of grime really started to grate on me. Becaus

Read more