Fill std error for one curve

I have this (X,Y) dataset and I want to plot with the std.error marked as shading. However, I have have problem when the std is 0. Here is a simple code is use for testing:

x = Data(:,1);
y = Data(:,2);
dy = 0.3*y;
fill([x;flipud(x)],[y-dy;flipud(y+dy)],[.2 .9 .9],'linestyle','none');
line(x,y);
alpha(0.1);

The dy is my std. error which is working fine, but when I change this to 0 it is looks strange (it should be equal the line)

Here is my dataset:

Data = [
  260.0000         0
  259.5000   -0.0166
  259.0000   -0.0487
  258.5000   -0.0445
  258.0000   -0.0437
  257.5000   -0.0638
  257.0000   -0.0583
  256.5000   -0.0880
  256.0000   -0.0961
  255.5000   -0.0706
  255.0000   -0.0863
  254.5000   -0.1051
  254.0000   -0.1140
  253.5000   -0.1329
  253.0000   -0.1307
  252.5000   -0.1433
  252.0000   -0.1625
  251.5000   -0.1366
  251.0000   -0.1359
  250.5000   -0.1413
  250.0000   -0.1438
  249.5000   -0.1538
  249.0000   -0.1352
  248.5000   -0.1844
  248.0000   -0.2098
  247.5000   -0.2066
  247.0000   -0.2031
  246.5000   -0.2036
  246.0000   -0.2479
  245.5000   -0.2791
  245.0000   -0.3187
  244.5000   -0.3629
  244.0000   -0.4218
  243.5000   -0.5147
  200.0000    2.5618
  199.5000    3.9747
  199.0000    5.4836
  198.5000    7.0462
  198.0000    8.4347
  197.5000    9.7647
  197.0000   11.1262
  196.5000   12.2604
  196.0000   13.3529
  195.5000   14.4072
  195.0000   15.3222
  194.5000   16.1851
  194.0000   16.9095
  193.5000   17.4813
  193.0000   17.8846
  192.5000   18.1166
  192.0000   18.3644
  191.5000   18.5597
  191.0000   18.5822
  190.5000   18.5643
  190.0000   18.3095
  189.5000   17.9620
  189.0000   17.6198
  188.5000   17.1708
  188.0000   16.8142
  187.5000   16.4826
  187.0000   16.1231
  186.5000   15.1229
  186.0000   14.6209
  185.5000   13.4553
  185.0000   12.5914
  184.5000   10.4794
  184.0000    8.5036
  183.5000    6.4980
  183.0000    4.7882
  182.5000    3.7990
  182.0000    3.4504
  181.5000    0.5280
  181.0000   -0.9536
  180.5000   -3.2450
  180.0000   -4.9457]

Answer 1

[Pawel Jastrzebski]

Look at you dy. I think you've changed it to dy = y. This is when you get a weird shape.

See line 80:

When dy = 0, you actually get a line:

[JB]

Hi Pawel. I am not sure what you mean. I you set the deviation (dy=0*y) then you should see no filling, right. I still get the strange fill when dy=0. Don’t you?

[Pawel Jastrzebski]

My mistake - I thought that by setting a dy = y I'm reproducing an appearance of the graph that you obtained for dy = 0 and my lazy eye saw it close enough. But now I can actually see that they are different.

Any way, you'll see from the second graph in my previous post that when I set the dy = 0 I get a straight line and no funny shapes. Even if I write the code your way, the output is the same:

I've done it in Matlab 2018a.

[JB]

That is strange. I still get the same wrong graph as showen above. I am using matlab 2015.

[Pawel Jastrzebski]

OK, try this. Maybe with this code you'll be able to track down what's plotting and how. The changes I've made are listed below, it's mostly cosmetics, but maybe Matlab will read this syntax better and produce the right plot for you:

• Moved out the x,y vales for the fill to a separate variables for readability, it's no xfill and yfill
• Used xfill and yfill to plot the chart using the plot function and dashed red line to see how it plots - this is the boundary of your fill plot
• added figure to initiate a new figure
• added hold on to keep all of your plots in a newly created figure
• swapped your line plot for plot plot
• added hold off at the end of the figure

x = Data(:,1);
y = Data(:,2);

dy = 0.3*y; % or dy = 0, it's the same thing

xfill = [x; flipud(x)];
yfill = [y-dy; flipud(y+dy)];

figure
plot(xfill,yfill,'--r')
hold on
fill(xfill,yfill,[1 0 0],'LineStyle','none');
plot(x,y,'k.');
alpha(0.1);
title('std.dev = 0.3')
hold off

The output:

Answer 2

[Patrick Haus]

Did you ever solve this?

I was trying to create the exact same error plotting routine and stumbled on this issue using Matlab 2015b. I happened to notice that it works as expected in Matlab 2022b. I had varied success in 2015b by using epsilon (eps) instead of 0 for your dy.