How can I force a MATLAB plot with two y-axes to have 0 at the same height for both axes?

I'm plotting two metrics vs a single x-axis. However, when the plots are output, the 0 for each of the y-axes is at a different vertical position in the graph. How can I cause them to be in the same position? (ie. at the bottom edge of the graph?) For reference, all values plotted are positive. See the attached:

Thanks!

Answer 1

[Walter Roberson]

To have the plots have a common origin, you need to adjust the ylim() for one or both plots such that the first value for the second plot has the same proportion relative to its ylim as the first value of the first plot has relative to its ylim.

rng(12345)
y1 = randn(50,1);
y2 = randn(50,1) .* sin(2*pi*rand(50,1));
yyaxis right
plot(y1)
ylim auto
yyaxis left
plot(y2)
ylim auto

[l1, u1] = bounds(y1);
[l2, u2] = bounds(y2);
[l1, u1; l2, u2]

ans = 2×2

-2.3876 2.1233 -2.4340 1.1394

%syms p
%(y2(1)-(l2-p))/(u2-(l2-p)) == (y1(1)-l1)/(u1-l1)
p = (y2(1) - l2 + ((l2 - u2)*(l1 - y1(1)))/(l1 - u1))/((l1 - y1(1))/(l1 - u1) - 1);
yyaxis right
ylim([l1, u1])
yyaxis left
ylim([l2-p, u2])

If my calculations are right, the origins of the two plots now align.

[Joseph Cunningham]

Hi Walter, and thank you for your answer. I'm assuming the graph you're showing was generateed by the above code. However, it looks like the y-axis zeroes are still not at the same height (ie. on the LHS 0 is at the top of the graph, while on the RHS 0 is in the middle of the axis). My hope is to have them both at the lower limit of the y-axis, as my values are positive.

With some modifications based on my code structure, I was able to apply your method, and the results are better, though not as exact as I hoped. Please see below for the results:

As you can see, when I zoom in on the graph, the RHS axis is still offset from the LHS axis zero. Do you have any suggestions on refining your original method to get a more exact match?

Thank you again for the help,

--Joseph

[Torsten]

x = 0:0.1:1;
y1 = x.^2 + 3 ;
y2 = 0.001*x.^2;
yyaxis right
plot(y1)
ylim ([0 4])
yyaxis left
plot(y2)
ylim auto

[Joseph Cunningham]

Thank you Torsten, this was the simple solution that I had somehow overlooked! Thanks!