We are at the end of the first month of My Year of Data…and I have learned a lot.  In the next few days I’ll post about some reflections on the first month, what works and what doesn’t and where I think this is going.   But first, a plot (click for embiggened version):

Withings scale plot

This is a plot of my daily weight from the start of my project until now.   Thanks to my Withings scale, weight is one of the few things that I have been able to consistently measure, every single day, without forgetting.   Overall, I’m quite excited to have lost about 5 pounds, but I am a bit concerned about the wild swings back and forth.   That’s what the holidays do to you, I guess!  My detox in mid-January seems to have regained my downward momentum, and I’m hoping here at the end of January I may be establishing a new baseline.

Withings Plot: iPad version

(an aside here about the Withings tool:  here is the same plot as rendered using the iPad app – again, click for larger version.  Note that this picture has some kind of ‘confidence bands’ which are actually a nice way of visualizing the overall trends and give a sense of the overall variance of the process, or for those geekoids out there, a quick view of the signal to noise ratio.)

Both of these plots show a trendline, which I think is just a simple moving average.  You’ll notice that the trendline lags behind the weight swings – which is exactly what you want it to do.  Fast weight loss (and gain, to a lesser extent) is not ‘real’, it is mostly temporary due to water weight or simply a stuffed belly.   It takes about a week at an established weight level for the trend line to ‘catch up’ — at that point you can more confidently say that you have stabilized at that weight.

One purpose of this project is to see how weight is correlates with other things that I track.  Obviously the most important link to weight is food intake.  I’ve struggled with being consistent with my food logging- it requires a dedication and diligence I cannot seem to keep up.   I was hoping to see how the foods that I eat related to other aspects of my life, and that is hard to do without consistent data.  But I did learn something very interesting thanks to my missing data!  It is not necessarily *which foods* I log, but rather *whether I log my food at all* which is a good indicator of whether I am losing weight or gaining weight.

The plot below shows the same data as the Withings chart above (I imported the data into R for more flexibility).  Here the curve is a standard smoothing curve, and so it does not have the lag.  The background pink color is shown on the days where I did regular food logging – the blank days were my lapses in logging.   If you look at this closely, you will see that when I am not logging my food, this is when the curve is going up (generally).  I think this is mostly psychological — if I know I am ‘off-the-wagon’ and I am eating and snacking more than I should, I tend not to log my food intake.  Conversely, if I am really dedicated to food logging on a given day, I am much more likely to eat healthy.

To me, this is at least a small piece of evidence in favor of my hypothesis…that the act of measuring something will affect what is being measured.    And although I am kind of disappointed in my inability to stay on the food logging plan consistently, it’s already helped me to learn something interesting about myself.

Weight chart with food logging days in color