Metrics interpretation

[BishKopt]

Hi, and thank for this great integration! :)

I have a couple of questions regarding the calculated metrics:

1. How to interpret "Body score"? How is it calculated?
2. How to interpret "Lean body mass"? In my case it's a bit off from simple "body weight" - "body fat weight" formula.
3. Should "Lean body mass" and "Visceral fat" be in kilograms (currently they don't have units)?
4. Why different metrics have different decimal precision (e.g. "Muscle mass" has 2, "Lean body mass" 1 and "Visceral fat" 0).

Thanks!

[r-jean-pierre]

Hi, I was close to create an issue, but it seems I'm not alone and I confirm that body mass and visceral fat do not have unit while others do.
I saw somewhere else someone asking for having BMI in kg/m². Personally, I'm not a big fan and I'm not sure if this unit is registered in the HA core, but, well, why not?
Now, thanks to the issue created by BishKopt, it came to my eyes that the decimal precision is indeed not the same. If this 'glitch' is not intentional and is easy to fix, I vote for having at least one decimal place of precision. Anyway, the sensors are not that precise, and the metrics are just rough approximate functions based on rough statistics gathered years ago from a non-fully representative population.

Thank for the integration and your consideration

[dbrgn]

I have a somewhat relateed question... Do these values make sense at all?

It makes sense that the body fat percentage can be estimated based on impedance, age, weight and gender. However, for example the in visceral fat calculation:

bodymiscale/custom_components/bodymiscale/metrics/weight.py

Lines 41 to 74 in 7294b9a

	def get_visceral_fat(
	config: Mapping[str, Any], metrics: Mapping[Metric, StateType]
	) -> float:
	"""Get Visceral Fat."""
	height = config[CONF_HEIGHT]
	weight = metrics[Metric.WEIGHT]
	age = metrics[Metric.AGE]
	if config[CONF_GENDER] == Gender.MALE:
	if height < weight * 1.6 + 63.0:
	vfal = age * 0.15 + (
	(weight * 305.0) / ((height * 0.0826 * height - height * 0.4) + 48.0)
	- 2.9
	)
	else:
	vfal = (
	age * 0.15
	+ (weight * (height * -0.0015 + 0.765) - height * 0.143)
	- 5.0
	)
	else:
	if weight <= height * 0.5 - 13.0:
	vfal = (
	age * 0.07
	+ (weight * (height * -0.0024 + 0.691) - height * 0.027)
	- 10.5
	)
	else:
	vfal = age * 0.07 + (
	(weight * 500.0) / ((height * 1.45 + height * 0.1158 * height) - 120.0)
	- 6.0
	)
	return check_value_constraints(vfal, 1, 50)

...all the input variables present are height, weight and age. If I'm not mistaken, this is essentially something similar to the BMI, but in no way is this value actually a measure of visceral fat, correct? Every person with the same age, height, weight and gender will get exactly the same visceral fat value, no matter how their body is built.

[Nickduino]

It's seems a bit disappointing but that's how Xiaomi calculates it.

Looking around at what exists, there doesn't seem to be any sort of consensus on a formula for VF that would use BF as a variable :-/