Calibration graph "Intercept" and "Raw value" axis values use mg/dl values and ignore mmol/l settings.

[sashasochka]

Hello, it would be more convenient to understand calibration values if they respected mmol/l settings on the Calibration Graph screen. Should also be very easy to implement. Would really appreciate.

[Navid200]

I am surprised.
I will be testing this when I return home.
If I don't get back to you in this thread in two weeks, please remind me.

[Navid200]

OK

This is what I see and understand.
The calibration graph has an x axis that corresponds to the raw values and a y axis that corresponds to the calibrated glucose readings.

If you change your units, the vertical (y) axis unit changes to correspond to the unit you have chosen.
The horizontal (x) axis units are independent of the chosen units because the raw values are always in mg/dL.

Please tell me if you have any questions.

[sashasochka]

I meant adding "mmol/l" or "mg/dl" to any number or respective label on the chart to avoid any source of potential confusion. E.g. to the "Raw value" label at the bottom of the calibration chart. Well, I had some quirky sensors lately and had a tough time calibrating them so I was looking at the calibration graph pretty often trying to understand the curve fit and it was a bit inconvenient having to do the conversions in my mind. Is there any other use case why anyone would look at the calibration chart? (And not prefer the values to be in mmol/l if that's their unit of choice)

[Navid200]

Is there any other use case why anyone would look at the calibration chart? (And not prefer the values to be in mmol/l if that's their unit of choice)

The unit of the vertical axis is mmol/L if that's the unit you have chosen. Is that not your experience?

[sashasochka]

Yes, it is. Unlike american (mg/dl) users who would have both axes using the same unit, it is not as easy to understand the chart, because the horizontal axis "Raw Values" uses mg/dl. What I'm also saying is that the vertical axis label explicitly specifies its unit of measurement, while the horizontal axis doesn't.

If the user changes their unit to mmol/L and erroneously think that the intercept limit is the same figure, it would be bad if they force the intercept.

I guess I do not entirely understand this sentence and why would the user misunderstand the figures, if they were all using the same single set of units as provided in the settings?

[PhilTU]

arguably the x-axis doesn't have to have units, a device could send a 0 to 100 output to represent its output value, and you input the corresponding glucose value to derive the calibration in your units of choice.

[Navid200]

If I look at the calibration graph, I can see that xDrip uses a linear approximation for calibration.
I can see how effective the last calibration has been considering each calibration has a weight and considering this weight is affected by the age of the calibration.

The unit of the x axis is not a part of any of those. I mean if the unit and values of the x axis change, I can still get the same info from the graph.
The x axis really is an electrical quantity. Hence the need for calibration to convert it to glucose.

You should be able to just ignore the values on the x axis.

Those become very important if you are a developer and need to make a change. But, as a user, I don't understand why you care about them.

[mcoffey73]

Unsubscribe

…

On Sat, Aug 5, 2023, 09:47 Navid ***@***.***> wrote: I am surprised. I will be testing this when I return home. If I don't get back to you in this thread in two weeks, please remind me. — Reply to this email directly, view it on GitHub <#2983 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6S75VRJVFYYZQ2ETHHUSQDXTZFHBANCNFSM6AAAAAA3EN53MQ> . You are receiving this because you are subscribed to this thread.Message ID: <NightscoutFoundation/xDrip/repo-discussions/2983/comments/6645783@ github.com>