Analytics of My Life
About six months ago I came across a post on Stephen Wolfram's blog where he collected data about his life for many years, then presented it in a bunch of graphs using Mathematica. It is very interesting if you want to check it out; however I wanted to do something similar. The problem is that I am not the CEO of a research company, and I haven't had software collecting data about me for years. I wasn't about to give up, so I downloaded some apps, did some research, and wrote some AutoHotKey scripts. Six months later I now have some data collected, and after learning a lot abut Excel I was able to make some visual representations.
Probably the simplest data I collected was about sleep using SleepBot. SleepBot's website makes graphs for you about your sleeping patterns. Here is the Sleep Distribution Graph:
It's fairly obvious to tell where the weekends are and where the summer is (I'm a student). The wake up times are surprisingly consistent; however the sleep times are not. Another graph shows the sleep duration, and that is all over the place. The drop near the end of the summer is traveling, and there are a few times where I had to get up early or stay up later than usual.
Here is a graph of the times I went to sleep. I usually went to bed around midnight during school, and around 2-3 during the summer. Obviously it's not perfect; there are some times when I stayed up later or went to bed earlier if I was really tired. There are few instances of going to bed after 3 or before 11.
The wake times are a lot more interesting. I have to get up between 7 and 8 for school, but I rarely have any other commitments to get up early for. If I don't have to get up at 7:30 usually I'll sleep in (hence the upward trend toward 11:00 excluding 7:00). Again I rarely get up before 7 or after 11.
According to the statistics I average 7.97 hours per night (I honestly have no idea how it is that much; probably sleeping in a lot on the weekends).
The first visualization Stephen Wolfram's post is a scatter plot of his email times. I don't send enough email to really see any trends of computer use, so I wrote a script that records the date and time in a file ever five minutes that that mouse is in a different position. I works pretty well; here is a scatter plot:
(Measured in five minute invervals)
Each dot represents five minutes of computer use. It's pretty obvious where the weekends are and where the summer starts. I usually use my laptop during class, but this is only for my desktop. You can see where I was on vacation during spring and summer break. The few dots in the middle of spring break are actually me using VNC to remotely connect to my computer. This isn't 100% accurate either. If a program (such as a backup) runs in the middle of the night it creates a dot (such as the ones at 5am every week). There is a gap around 6pm when I eat dinner, and around 2pm when I eat lunch over the summer.
Here is a similar graph of my browsing pages for the summer of 2013:
It's not as much data, but you can see where I went on a trip in late July (there is some activity in the evening since this included my laptop). The very late/early dots in early June are when I couldn't sleep. It's funny to look at the actual pages. At about 4am they are "how to fall asleep" and at about 6:30 they are "negative effects of staying up all night". Also, getting on your computer is about the worst thing you can do since it emits blue light, but since it was summer I could sleep until 1pm if I needed to.
This is a graph of the time where I use my computer. It drops around 6-7 pm, and also around 2pm for lunch. I thought it might look different over the summer, so this is a graph of just the summer data:
The prime time for using the computer shifts a little later, and there is definitely a more prominent break around 6-8pm (coolest time to go outside).
Here is a graph of computer use by day of the week:
It doesn't vary as much as I though it would between the week and weekends, but I guess I do more activities off the computer during the weekend.
Here is a graph of computer use by month. June is the biggest by far. I traveled a lot on July, and school started again in August. February has incomplete data.
According to Rescuetime I have spent 578 hours on the computer since April, which is about 4 hours per day. I have honestly spent 94 hours of that on Reddit, 92 on Youtube, 43 on JavaW (looks like Java devleopment, but most of it is probably Minecraft), 33 on Google, and 20 hours on MS word (more on my laptop though). I have spend a lot more time browsing the internet and reading smaller news websites that don't appear on here.
Despite spending 94 hours on Reddit, I actually do a lot of studying:
I averaged studying 2:48 excluding 0:00 days such as Friday and Saturday, and 2:10 including them. It seems that there is almost always a dip in the middle of the week, and I can't identify the reason. I usually had pretty even classes throughout the week.
Here is a scatterplot of Reddit use:
(Program graphs are in 1 minute invervals)
The more times I just check Reddit for just a minute and not 30 the better. These dots are only a minute so it's not as bad as it looks.
I apparently don't browse Reddit that much at 2pm or late at night. I don't know why I don't at 2:00, but I am probably working on something late and if I didn't know what to do I would go to bed instead of going on Reddit.
The time is pretty even throughout the week except for Saturday. I am probably doing stuff Friday night and doing homework Sunday.
Probably the worst graph I made was this:
I don't think I need to explain anything.
I could make this graphs for almost any program/website, but you are probably getting bored so here's one more that I thought was interesting.
I don't know why I don't listen to much music between 6 and 7pm, but the evening/night seems to be the biggest period. (Foobar2000 is a music player that you should get if you haven't already)
I don't know why I don't listen to music on Wednesday either.
Other miscellaneous things I have collected are below:
Whatpulse has said I have typed 2.97 million keystrokes in the past ~year, and clicked about 1 million times. My record for most keystrokes is about 60,000, which is two hours of typing at 100wpm (I can type that fast). I have averaged about 9540 keys this year, and 3193 mouse clicks. I have downloaded almost a Terabyte, and uploaded about 417 Gigabytes.
I have a map of everywhere I've been, but for obvious reasons I won't show it. Since about Thanksgiving 2013 I have traveled about 11,000 miles. Google collects that data, and it can be found at http://maps.google.com/locationhistory. I don't ever recall agreeing to it, and I usually check every "do not track" box possible. (Before somebody calls me a hypocrite: collecting useless data about myself with a script I wrote and storing it locally is different than a company/government tracking me).
I think this sort of data can generally help you be more productive and see exactly how and when you spend your time. It's not that hard to collect digitally; it took me about an hour to write this script. I really would want to see more data about non-digital stuff, but that would be a lot harder to collect and have more privacy concerns. Anyway, thank you for reading this, and I hope to see more people try similar experiments.