Going geeky with automated diet devices

LAT1963

I have a Withings scale that automatically uploads my weight every time I use it, and from which a list of weights with date/time stamps can be downloaded to Excel.

So, putting my geek on, I downloaded Feb-April to an Excel file, inserted a graph with x = date and y = weight, and then (because it looked pretty linear), did a linear curve fit with 'show equation' checked in the properties tab for the graph. You have to muck around a bit to get the dates to show as dates instead of Excel's numerical internal representation of dates on the x-axis--I forget how I fixed this but its something you'll encounter.

Why be so geeky? Because I'm attending someone's wedding this June and wanted a realistic estimate of what my weight will be (formula says 215 lbs) so that I can buy an appropriately-sized dress. (I already own an appropriate dress that fits me at about 200 lbs but now I know I'll be too big for it so I'll need to buy another).

I've had the equation for a week and it seems to be predicting my progress pretty accurately (I went over target 3 times this past week and my weight is only 0.2 lbs above predicted, but that's to be expected due to my incomplete compliance with program).

Anyway I wanted to share the news that this kind of analysis seems to work, in case anyone else has an event coming up and needs a reasonable estimate of their likely size.



The semi-isolated sawtooth of weights about 2 lbs above the main trajectory of the data is due to weights taken at bedtime in addition to those taken first thing in the morning. I left them in for the curve fit because I didn't think they were numerous enough to make much difference and they do represent the likely scatter I might see.