Steppin’ Up Challenge, Part 3

Welcome to the third and final part of the Steppin Up Challenge analysis! If you missed part 1 & 2, I encourage you to check it out here and here. So, here we are, analyzing the results of the 20 competitors in this four week challenge where we tracked active minutes and steps taken on a daily basis.

Last time we looked at the scatterplot of active minutes versus steps taken across all competitors. We looked for competitor’s who strayed far away from the linear regression, thereby having a higher step/min count or a lower step/min count.

Scatterplot_distance

We looked closely at competitor P, who I renamed Powerhouse. We found some interesting conclusions. Our linear regression still provided fairly reasonable results. However, all of the conclusions I made that were based on my knowledge of her as a runner were totally debunked, because she didn’t run at all during this 4 week period. Oops! Perhaps I’ll have better luck analyzing my own data.

And I also promised that I would analyze my own data, because anything I could share about someone else’s data, I should also share about my own. So, what about competitor “N”?  N was dubbed Namaste because I know that competitor N does a lot of yoga and walking and not much else. Let’s take a look at my, I mean her, time series.

Namaste_timeseries_socialmath

Here there is a very clear relationship between steps and minutes. Which is actually a little surprising. If Namaste is doing a lot of yoga, I would expect her activity minutes to increase, but for her steps to stay flat. However, these time series are very much in phase with one another. The peaks in Namaste’s steps correlate well to her activity times. Let’s take a look at her scatterplot.

Namaste_scatter_socialmath

Namaste has an average of only 70.4 steps per active minute. This is lower than walking speed, which is reasonable given her choice of exercise (yoga and occasional biking). I would hedge my bets that the days which are furthest from the linear regression on lower right are the days that Namaste did yoga. The data points above the line are the days where she mostly walked everywhere.  In fact, I would bet that the minutes and steps are well correlated because yoga days also require extra walking (to get to the studio). However, I only know that because Namaste is my data. I wouldn’t be able to inuit this if it wasn’t my data.

On her, I mean my, fastest day, June 24th, she did 128 active steps/active minutes. Which is a little above average walking speed. Clearly, walking is fast enough for her. …I mean me. I don’t do anything faster than a walk. It’s a turtle’s life for me.

turtle_turtle        turtle_turtle       turtle_turtle

The only other competitor I really want to share with you is competitor M, who happens to be one of the winners of the whole competition. I’m going to show you his time series, and I want you to guess what happened:

M_timeseries

Do you have a guess as to what caused the low numbers of steps and minutes for 4 days in late June/early July?  To me, it was clear, competitor M got sick. I talked with him about it and yes, he over did extended himself in week 1 and got sick during week 2.  However, his exercise level was high enough that he made up for it and was one of two winners of the competition. So, sometimes, the data tells a clear story. But sometimes it doesn’t…

I’m not sure what the conclusion is here. Certainly, there are lots of interesting things I can learn about your life if I have access to daily data related to your lifestyle. Data sleuthing can intuit a fair number of things about your life and fitness level. However, with nothing but steps and activity time, there might not be enough information to make any real conclusions. Maybe if I had GPS data I could really get somewhere! But, with this information I can’t even tell you how fast Powerhouse can run. And I can’t confidently tell you what days Namaste did yoga. But I can make some basic observations about your life, as shown here.

Did I cross the line between comforting or creepy? Only you can be the judge. But you should know, that FitBit or Jawbone or whatever your pedometer of choice is… they have this data. Your cellphone carrier might have it too. You may also be giving it to Niantic if you play Pokemon GO. And, of course, PRISM probably has it. The New York Times recently did a piece on the Chinese app, WeChat, and how it is changing the way business is done. …and the government gets all the data. In the US we tend to partition our data across companies so different companies each only have part of your data (like what I did here). And this usually keeps one entity from having all the data. But what would happen if such a monster app existed in the US?

Advertisements

About Samantha from SocialMath

Applied Mathematician and writer of socialmathematics.net.
This entry was posted in Communicating Math, Exercise and tagged , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s