How do you graph your weight?

dolorsit

This may seem like such a simple question, but I'm curious if there's a better way. Since your true weight is really some function of what you've eaten the last few days and what you've expelled, you can never really know this, but only approximate it. The report on MFP doesn't perform any sort of curve fitting of your weight, but various other sites and apps seem to have better approaches. One of the ones I put my data in is trendweight, but since it uses a trailing average, it seems to be way behind all the time and is probably only accurate if you're maintaining weight.

So, if I want to answer questions like how much weight I lost in the last 7/14/30 days, what would you suggest would be a good way to do it?

FWIW, this is what trendweight looks like for me:


Or perhaps I'm overthinking this?

cwolfman13

I'm going to go with overthinking...

Lietchi

I have several 'weights' in my excel spreadsheet:
- individual weigh-ins
- average weight over 15 days (7 days before to 7 days after)

My 15 day average also goes up and down, but less than the individual weigh-ins.

The individual weigh-ins are interesting when used in a 'stacked' manner to compare between months:

dolorsit

cwolfman13 wrote: »

I'm going to go with overthinking...

I am definitely prone to that!

cmriverside

@Lietchi
You win for best use of weight data in a graph.

dolorsit

Lietchi wrote: »

The individual weigh-ins are interesting when used in a 'stacked' manner to compare between months:

That is very interesting. I've only been on this latest diet since Jan 1, so don't have months of tracking yet, but maybe I could do that by week. Thanks for the suggestion.

AnnPT77

dolorsit wrote: »

This may seem like such a simple question, but I'm curious if there's a better way. Since your true weight is really some function of what you've eaten the last few days and what you've expelled, you can never really know this, but only approximate it. The report on MFP doesn't perform any sort of curve fitting of your weight, but various other sites and apps seem to have better approaches. One of the ones I put my data in is trendweight, but since it uses a trailing average, it seems to be way behind all the time and is probably only accurate if you're maintaining weight.

So, if I want to answer questions like how much weight I lost in the last 7/14/30 days, what would you suggest would be a good way to do it?

FWIW, this is what trendweight looks like for me:


Or perhaps I'm overthinking this?

I hear what you're saying about interesting metrics and charts; I'm a data geek myself. This is just an alternative conceptual frame to consider as an adjunct to trend watching:

Personally, I think I don't have a "true weight". I have a current range, and a long-term trend.

Current range: Over the past 7 days, my daily morning weigh-ins have ranged from a low of 123.8 pounds, to a high of 125.8 pounds (and not in linear fashion, either - ups and downs). I haven't recorded weights at other times of day, but those would make the range even wider.

I think I have a long-term trend, over weeks to months. I do use a weight-trending app (Libra, in my case). But I can also see the trend if I compare past and current weight ranges.

For example, in the same week in January, the range was 124.6 to 126.8. In December same week, 124.8 to 128.4. Last February (2020), 134.4 to 136.6.

To me, this suggests that I'm losing weight ultra, ultra slowly, less than a pound a month, practically invisible . . . which is my plan and intention, at this point in year 5 of weight maintenance.

In this conceptual frame, the only answer to "how much weight have I lost" is an approximation, in my case, about 10 pounds in the past year. It's not an exact number.

ahoy_m8

cmriverside wrote: »

@Lietchi
You win for best use of weight data in a graph.

Agreed!

I just use daily weight and 7-day trailing average. BUT, I do superimpose daily intake and 7-day trailing average on the same graph, so it's clear how a spike in one is followed closely by a spike in the other. Hopefully both smooth out on the trailing average.

Lietchi

cmriverside wrote: »

@Lietchi
You win for best use of weight data in a graph.

I wish I could take credit for it, but it was actually another MFP'er who posted a similar graph and I thought it was a great idea. I love my spreadsheet and graphs 😆

gcminton

dolorsit wrote: »

This may seem like such a simple question, but I'm curious if there's a better way. Since your true weight is really some function of what you've eaten the last few days and what you've expelled, you can never really know this, but only approximate it. The report on MFP doesn't perform any sort of curve fitting of your weight, but various other sites and apps seem to have better approaches. One of the ones I put my data in is trendweight, but since it uses a trailing average, it seems to be way behind all the time and is probably only accurate if you're maintaining weight.

So, if I want to answer questions like how much weight I lost in the last 7/14/30 days, what would you suggest would be a good way to do it?

Or perhaps I'm overthinking this?

Well, since you know you can only approximate it, perhaps the best way to proceed is to be okay with an approximate answer. I also use TrendWeight's website since it syncs with my scale. Either the Actual readings or Trend readings will be perfectly good for telling me what I've lost for a set time period, as long as I'm comparing Actual to Actual or Trend to Trend. My experience with other trending apps or websites has worked the same way as TrendWeight, and the trailing average you mention is exactly what it's supposed to do; it's smoothing out the spikes to filter out some of the noise.

If you'd rather focus on your literal scale readings, you can still do that just as easily.

dolorsit

Ok. I'm just trying to assure myself that I'm not losing weight too quickly. Since I started at the beginning of the year, I'm losing on average 1.2kg/week which is probably a little fast but I'm doing a lot of running and cycling. I've slowed down over the last week but trendweight is taking a while to catch up since it lags so much. I guess I'd like a spreadsheet column that lights up red to tell me to slow down...

I do have another spreadsheet graph that plots my 5 day average average of weight from scale and weight from body fat% assuming constant lean body mass. It looks like this: