How to shift experimental data (not a function) in a loglog plot?

3 views (last 30 days)
I am trying to shift data just for viewing purposes, the actual values would not be relevant. I just don't want the data to overlap. Since it is on a logartithmic scale, it is not as direct as adding a factor to each value. And all recommendations I see involve shifting the mathematical log function. I wonder if there is a way to evenly shift experimental data on a loglog scale.
clear; clc; clf; close all;
x = [0.015 0.02 0.05 0.1 0.2 0.3 0.4 0.5];
y1 = [1.1302e-11 1.9874e-11 3.5178e-11 8.0091e-11 1.6816e-10 2.3635e-10 3.3775e-10 4.1115e-10];
y1fit = [1.2258e-11 1.6337e-11 4.0783e-11 8.1477e-11 1.6278e-10 2.4401e-10 3.252e-10 4.0636e-10];
y2 = [7.3093e-11 8.5947e-11 1.5178e-10 2.3051e-10 3.121e-10 4.0186e-10 4.8517e-10 5.8097e-10];
y2fit = [7.3686e-11 8.6974e-11 1.4748e-10 2.199e-10 3.2787e-10 4.1418e-10 4.8887e-10 5.5596e-10];
y3 = [2.2116e-11 3.3031e-11 8.0283e-11 1.3912e-10 2.7158e-10 4.16e-10 5.8192e-10 7.2361e-10];
y3fit = [2.3459e-11 3.0991e-11 7.5229e-11 1.4714e-10 2.8781e-10 4.2612e-10 5.6293e-10 6.9864e-10];
figure
a = loglog(x, y1,'ko', 'MarkerFaceColor', 'b', 'MarkerSize',10);
hold on
a1 = loglog(x, y1fit, 'k--', 'LineWidth', 2);
b = loglog(x, y2,'kd', 'MarkerFaceColor', 'm', 'MarkerSize',10);
b1 = loglog(x, y2fit, 'k--','LineWidth', 2);
c = loglog(x, y3,'ks', 'MarkerFaceColor', 'r', 'MarkerSize',10);
c1 = loglog(x, y3fit, 'k--', 'LineWidth', 2);
I edited it in paint; getting something like this would be the goal, no overlapping data and can easily view the different slopes (I don't mind the higher y value).

Accepted Answer

Les Beckham
Les Beckham on 5 Jan 2023
Edited: Les Beckham on 5 Jan 2023
Instead of adding an offset, multiply to shift the data. I used a factor of 10 as an example.
x = [0.015 0.02 0.05 0.1 0.2 0.3 0.4 0.5];
y1 = [1.1302e-11 1.9874e-11 3.5178e-11 8.0091e-11 1.6816e-10 2.3635e-10 3.3775e-10 4.1115e-10];
y1fit = [1.2258e-11 1.6337e-11 4.0783e-11 8.1477e-11 1.6278e-10 2.4401e-10 3.252e-10 4.0636e-10];
y2 = [7.3093e-11 8.5947e-11 1.5178e-10 2.3051e-10 3.121e-10 4.0186e-10 4.8517e-10 5.8097e-10];
y2fit = [7.3686e-11 8.6974e-11 1.4748e-10 2.199e-10 3.2787e-10 4.1418e-10 4.8887e-10 5.5596e-10];
y3 = [2.2116e-11 3.3031e-11 8.0283e-11 1.3912e-10 2.7158e-10 4.16e-10 5.8192e-10 7.2361e-10];
y3fit = [2.3459e-11 3.0991e-11 7.5229e-11 1.4714e-10 2.8781e-10 4.2612e-10 5.6293e-10 6.9864e-10];
figure
a = loglog(x, y1,'ko', 'MarkerFaceColor', 'b', 'MarkerSize',10);
hold on
a1 = loglog(x, y1fit, 'k--', 'LineWidth', 2);
b = loglog(x, 10*y2,'kd', 'MarkerFaceColor', 'm', 'MarkerSize',10); % <<< multiply to shift up
b1 = loglog(x, 10*y2fit, 'k--','LineWidth', 2); % <<< multiply to shift up
c = loglog(x, y3,'ks', 'MarkerFaceColor', 'r', 'MarkerSize',10);
c1 = loglog(x, y3fit, 'k--', 'LineWidth', 2);
grid on

More Answers (2)

William Rose
William Rose on 5 Jan 2023
@Alfredo Scigliani, Multiply all y-values of the trace in quesiton by a constant. This will cause a parallel vertical shift on a log-log plot.
  1 Comment
William Rose
William Rose on 5 Jan 2023
x=10.^(sort(2*rand(1,8))-2);
y1=x/10;
y2=x.^2;
subplot(211)
loglog(x,y1,'-ro',x,y2,'-bx');
xlim([.01,1]); title('Before shift');
legend ('y1','y2','location','southeast')
subplot(212)
loglog(x,10*y1,'-ro',x,y2,'-bx'); %parallel shift of y1
xlim([.01,1]); title('After shift');
legend ('shifted y1','y2','location','southeast')
Good luck

Sign in to comment.


Fifteen12
Fifteen12 on 5 Jan 2023
I believe you have to do it point by point. Since a 1 unit increase is equivalen to 10x the value, if you wanted to shift upwards by 1 unit you would have to times by 10. For example:
x = [0.015 0.02 0.05 0.1 0.2 0.3 0.4 0.5];
y = [1.2258e-11 1.6337e-11 4.0783e-11 8.1477e-11 1.6278e-10 2.4401e-10 3.252e-10 4.0636e-10];
yshifted = y * 10;
loglog(x, y, 'r', 'LineWidth', 2)
hold on
loglog(x, yshifted, 'b', 'LineWidth', 2)
You can change the shift value from 10 to any other digit as well.

Products


Release

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!