Health metric forecasting

ajdavis2004

Why has no one yet joined the new group I started called "Health Metric Forecasting"?

While working to achieve my fitness goals, I'm also trying to verify my fitness model using rigorous resolvable metrics. Like any other scientific endeavor, my fitness model should be verifiable by quantifying both the desired outcome as well as the time-weighted average prediction error. Has no one joined me because this is just another foolish idea of mine?

claireychn074

I for one have no desire to publicly post detailed data about myself online. Good luck getting many people to share that stuff - especially as you’re talking about some kind of model. Where’s your DP statement / PIA? How are you anonymising the data, where are you holding it, what’s it being used for etc?

AnnPT77

I don't think your effort is foolish, but I think it is unlikely to be shared by many others, at least here. That's not a criticism.

I'd say I'm somewhat on the data geek side compared to the average person, and I'm a definite fan of data-driven decision making, but I don't really see the point (for me personally) of a complex statistical model.

It's not the most common, but we do see people here pretty often struggling with issues like calculating portion size that will fit their calories. I don't think they will want to pursue rigorous predictive models and metrics either, probably. (I'm sure some would like a formula where they could plug in values and get a sure-to-be-correct result, but I think much of success in the health/fitness realm is very individual, and I'm pretty sure that people do poorly when trying to apply statistical models they don't understand . . . at least there seems to be a fair amount of difficulty here understanding the models underlying exercise calorie estimating, BMR/RMR or NEAT or TDEE estimating, or even weighted-average weight trend models.)

Speaking personally, the validation for my efforts is that I reached a healthy weight over 8 years ago and have stayed there since (after previous years of overweight/obesity); that my multi-week average body weight changes are quite predictable based on the calories I log (food and exercise), and have been since I developed a personalized calorie needs estimate using my own data (since it turns out my needs are surprisingly non-standard); that the physical fitness metrics I care about are typically at least in the upper quartile for my demographic segment (talking about things like pace/distance relationship in my sport, resting heart rate, heart rate recovery, performance on certain test activities, etc.); that my blood test results and other health markers range from normal to the extra-good side of normal; that my body fat percent is substantially lower than average for my demographic; etc.

I can see why a more nuanced model would be fun for someone who likes modeling.

But for me, the much less complicated "model" I've used has gotten me to my goals and kept me there with much less complexity, record keeping, and effort. (That would be logged calories in MFP, various fitness metrics primarily from my fitness tracker but also a few other routine sources, a weight-trending app, and similar stuff.)

Getting more statistically nuanced wouldn't be fun for me, and I think it's unlikely to deliver sufficiently better results that the extra effort (or fun of modeling) would be worth it for me. There are other ways I'd have more fun spending my time. YMMV, and that's what makes interacting with others fascinating, if you ask me.

What you're doing is interesting, and if it suits your style, that's great, sincerely. I wish you success, and I hope you find some happy participants . . . but I have doubts that many people here would be excited about the experiment. You might have better luck in group for advanced-degree personal trainers, statisticians who are pursuing health/fitness improvement, or something like that.

ajdavis2004

OKAY, thanks for the comment - here is my personal privacy statement. I am not a Federal agency nor do I have any commercial interest in your data. I am solely interested in ascertaining the validity of the fitness models we are currently, or have been, using for your forecast with the sole intention of improving our models.

"I only have access to/collect Personal Data that you voluntarily give me via the myfitnesspal.com community, email or other direct contact from you. I will not sell or rent this Personal Data to anyone. I will use your Personal Data to respond to you, regarding the reason you contacted me."

Section 208 of the E-Government Act requires all Federal agencies to conduct a PIA when developing or procuring new information technology involving the collection, maintenance, or dissemination of information in identifiable form or when making substantial changes to existing information technology that manages information in identifiable form.

Good luck with your fitness model. I hope you find some sort of validation for your efforts somewhere.

Corina1143

You could have just said "I'm trying to lose 17 pounds in January." I'm sure you would have received many replies. Perhaps not the kind you wanted?