Scale right y-axis to the values on the left y-axis

Hello everybody,

I'm trying to plot 3 different arrays in a 2D figure:

figure(100)

yyaxis left

plot(IR_range/1000,angle)

yyaxis right

plot(IR_range/1000,distance/1000)

Both plots look remotely the same, but not quiet:

As you can see both y axes are almost proportional to one another, but not 100%. I'd like to make the orange curve fit the blue curve (into one curve) and let the right y-axis (orange) scale change accordingly. Anybody that can help me

Answer 1

[Cris LaPierre]

I don't see how changing the scale of your Y Axis is going to help you. No amount of uniform stretching or compressing is going to make these two curves overlap. For the first half, your orange curve is above the blue, but for the second half, it is below. This means you will have to amply a non-uniform transformation to get the curves to overlap.

In the process, you are then modifying your data. So in the end, you might end up with curves that overlap, but no longer represent the event they were recorded from.

[Noah Stam]

Indeed a non uniform transformation would be needed. In that sense the label ticks on the right side y axis would be more like data tips that indicate the specific value only at that point. Since the data on the left and right y axes are so closely related, I didnt want to use a colored scater plot, like its a completely different parameter.

[Noah Stam]

The scale between every pair of lable ticks in the right side would obviously be different (and not even linear for that matter).

[Noah Stam]

Maybe even need to use additional ticks that indicate the scale distribution between two 'main' right y axis label ticks.

[Cris LaPierre]

While you can place your ticks anywhere you want, axes must have a uniform scale. This means to achieve what you want, you will have to modify the actual x and/or y data values. That makes this a data manipulation question, and not something related to axis scaling.

[Noah Stam]

Okay, fair enough. Thanks for the help! I guess ill have to find another way to represent this data in a nice way then.