Compare live data with recorded data sets

Hello!

Not too much of a newbie but I can’t wrap my head on where to start.

I have 4 sets of prerecorded data curves that are in excel.
Is it possible to compare live stream data from the arduino to the 4 data sets?

Basically I want the arduino to compare the livestream data to match it to the closest prerecorded data curve. Once it’s matched , it will output a certain number on the lcd screen coorisponding to that matched curve

Any suggestions would be awesome . Complete base code will be highly appreciated

Thank you all

Add error bars to your prerecorded data points, for example plus or minus 5%.
For your measured data, if a value falls within that error bar then it belongs to that dataset.

For more sophisticated matching you could use weightings so a measurement that is within 1% scores higher than one that is within 4%.

As @Delta_G said, the question is very vague.

Why would you do this on the Arduino?

Seems that the computer running Excel would be better placed...? :thinking:

You're going to have to think about how you want to define "match" here.

eg, is it just the "shape" of the curve matches? what if your "live" data matches the beginning of one of the recorded sets, and the end of another?

One of the easiest ways to match curves is to plot them all on the same graph ...

I can imagine that, but you need to provide really basic information about the problem, like examples of the data, what you mean by "match", whether time-shifting is required, etc.

1 Like

Then you should probably be asking here:

Thanks for the reply everyone.

So each recorded data set has totally different resistance values in a log curve.

Yes a 5 or 10% tolerance of the curve is acceptable .

The program should easily be able to decern the correct curve to match to.

Are there other better programs to do this task?

I’ll try the service if need be, thank you for the suggestion, never knew it existed. Is it expensive?

:point_up_2: This! :point_up_2:

Whether you do it in Excel, or some mathematical analysis package, or write yourself some code - I think the PC is the place to do it!

Python is well-equipped and widely used for data analysis tasks like this

https://matplotlib.org/

etc ...

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.