shade area between graphs

I want to shade the area between these graphs

but this certainly does not do the trick:

fill([t t(end:-1:1)],[y1 y2(end:-1:1)],'y')

I also downloaded and tried "jblill"

http://nl.mathworks.com/matlabcentral/fileexchange/13188-shade-area-between-two-curves

but that did not seem to amend the outcome so far

[Huijian Huang]

have you solve this problem at the end? I am also struggling with similar situation

[Image Analyst]

Yes, there were answers below, one of which he accepted though both should work.

Answer 1

[Image Analyst]

Try using fill() instead of area():

x = 1 : 300;
curve1 = log(x);
curve2 = 2*log(x);
plot(x, curve1, 'r', 'LineWidth', 2);
hold on;
plot(x, curve2, 'b', 'LineWidth', 2);
x2 = [x, fliplr(x)];
inBetween = [curve1, fliplr(curve2)];
fill(x2, inBetween, 'g');

[Michiel]

gives the exact same plot as I posted above (although I used the graph from my original post this time, which is arbitrary)

[Image Analyst]

Even if I make data that cris-crosses, it seems to work:

% Create sample data.
x = 1 : 100;
curve1 = log(x);
curve2 = log(x) + rand(1, length(x)) - 0.5;
% Plot it.
plot(x, curve1, 'r', 'LineWidth', 2);
hold on;
plot(x, curve2, 'b', 'LineWidth', 2);
x2 = [x, fliplr(x)];
inBetween = [curve1, fliplr(curve2)];
fill(x2, inBetween, 'g');
grid on;

Michiel, please attach a few of your data files so we can do it with your actual data.

[Michiel]

I found the problem... I used a column vector instead of a row vector... had to transpose the damn thing... rookie mistake 8)

[Image Analyst]

OK. Good, so fill() worked. Can you mark my Answer as Accepted then? By the way, for column vectors, use semicolon instead of comma and flipud() instead of fliplr():

% Create sample data as column vectors.
x = [1 : 100]';
curve1 = log(x);
curve2 = log(x) + rand(length(x), 1) - 0.5;
% Plot it.
plot(x, curve1, 'r', 'LineWidth', 2);
hold on;
plot(x, curve2, 'b', 'LineWidth', 2);
% For column vectors, use flipud(), for row vectors use fliplr().
x2 = [x; flipud(x)]; % Use ; instead of ,
inBetween = [curve1; flipud(curve2)]; % Use ; instead of ,
fill(x2, inBetween, 'g');
grid on;

[Hassan Raheem]

Hi, if I have to curves , but the data for each curve differ from the another (the length is not the same) , How can I plot them? Thank you EX: x1 = read from csv file1 first column x2= read from csv file2 first column y1= read from csv file1 2nd col. y2= read from csv file2 2nd col plot(x1,y1) hold on plot (x2,y2) Then how can I shade the area between these two curves?

[Julien Reynes]

It is exactly what a wanna do, but my problem is that my two curves contain NaN (they are not continuous) So if I plot, no filling appears, just boundaries. Any idea how to solve it easily?

[Tom]

If you have data sets of different lengths in the same figure, I suggest interpolating both datasets in such a way that their x-axes are identical. Since the PLOT function implicitly interpolates the data linearly, this will not change the way your data is displayed.

If you have data that contains NaNs, what do you want to achieve? How should the filling be? Should it stop at discontinuities? Should it act as if the discontinuities aren't there?

[Tilen Thaler]

Thank you. Exactly what I was looking for.

[Rogelio Garcia]

Does this method work when both curves have different array lenght?

[MadjeKoe]

Does anybody maybe know how to remove the outside lines and to only plot the shade? Thank you in advance

[Image Analyst]

@MadjeKoe simply don't call plot() and only call fill().

[Jason Thomas]

Hoping this post from 9 years ago is still monitored.

I tried the suggestion above, but instead of filling the area green, it just added thin black lines at the upper and lower bounds. What am I missing?

plot(averageFF2W5I(:,2),averageFF2W5I(:,1), 'k', 'LineWidth', 2);
plot(averageFF2W5I(:,2),averageFF2W5I(:,[8 9]), 'k--', 'LineWidth', 2);
fill([averageFF2W5I(:,2), fliplr(averageFF2W5I(:,2))], [averageFF2W5I(:,8), fliplr(averageFF2W5I(:,9))], 'g');

averageFF2W5I(:,2) = Strain Values

averageFF2W5I(:,1) = Stress Sample Mean

averageFF2W5I(:,[8 9]) = 95% Confidence Interval

[William Rose]

@Jason Thomas,

I don't have your array, so I will use analogous data.

x=0:20; y=x.^.5; ci1=0.8*y; ci2=1.2*y;
fill([x, fliplr(x)], [ci1, fliplr(ci2)], 'g', EdgeColor='none');
hold on
plot(x,y, 'k', 'LineWidth', 2)
plot(x,[ci1;ci2], 'k--', 'LineWidth', 2)

I moved fill() before plot(), so that the line in the middle would not get covered by the fill. Alternatively, you could specify a semitransparent fill. I specified no edge for the fill. I added hold on after fill().

Answer 2

[Star Strider]

Try this:

x = linspace(0,10*pi);
y1 = sin(x);
y2 = cos(x);

figure(1)
plot(x, y1)
hold on
plot(x, y2)
patch([x fliplr(x)], [y1 fliplr(y2)], 'g')
hold off

Experiment with your data to get the result you want.

[Michiel]

ttom = time vector [1:variable_interval:20]

y1 = the red line

y2 = the blue line

patch([ttom fliplr(ttom)], [y1 fliplr(y2)], 'g')

I had this before as well, somehow I'm doing it wrong

[Star Strider]

I would have to have your data to find out what the problem is. (If you want to attach it using the ‘paperclip’ icon, a .mat file is best.)

Since I don’t have your data, add a NaN value at the end of each of your data arrays (both x and y variables for both data sets) and see if that makes a difference. Just guessing.

[Star Strider]

If you would share your data, I could likely provide a definitive solution. I have no idea why your data are not working with my code.

[Star Strider]

Did my idea work with your transposed vectors? Mine requires row vectors as written here. I perhaps should have specified that.

[DETELINA IVANOVA]

I am using this last suggestion by Star Strider to plot error bar shading using patch, but in some cases I get in addition to the shading straight lines (see the third plot at the bottom in the attached figure) I have eliminated the points with NaNs. It seems ok in the other two cases (the upper plots in the same figure). What can be causing this?

[James Stephenson]

This answer is too late for you, but maybe it will help others. The last point of your vector is your starting point as well, so the shaded area is doubling back to (0,0). That's why you're getting this wonkiness.

[Yash Mistry]

how can I avoid that?

[Star Strider]

@Jason Thomas —

If you have not already done so, check to see if there are NaN values in your data. One way to do that is:

NrNaNs = nnz(isnan(averageFF2W5I))

Your options then are to to remove them using the rmmissing function or interpolate them with the fillmissing function.

However, one problem is that you have column vectors, and you are using fliplr to flip a column vector. That is going to produce the same unchanged column vector, of course.

An improvement to your fill call could be:

fill([averageFF2W5I(:,2); flip(averageFF2W5I(:,2))], [averageFF2W5I(:,8); flip(averageFF2W5I(:,9))], 'g');

Note my use of the semicolon (;) vertical concatenation operator, and the generic flip function that works without further elaboration on vector arguments.

.

[Jason Thomas]

@Star Strider THANK YOU!

There were several NaN values in the confidence interval. Once I removed those with rmmissing and used your code with "flip" instead of "fliplr" it worked perfectly.

[Star Strider]

My pleasure!

A Vote would be appreciated!

Answer 3

[Javier Montalt Tordera]

This can be simplified using the function shade (in MATLAB File Exchange).

The syntax for the above problem would be:

shade(t,y1,t,y2,'FillType',[1 2;2 1]);

The FillType option specifies that the area between lines 1 and 2 should be filled, whether 1 is above 2 or the other way round.

[SEETHA JAYAN]

would you please help me ,

i used the above code to shade between min and maximum but I couldnt change edge color and face colur

[Javier Montalt Tordera]

You should be able to change line color and fill color using the properties 'Color' and 'FillColor', respectively, specified as name-value pairs.

[Giacomo Giuliani]

Is it possible to fill an area between three or more curves without covering none of the curves between the edges?

[Image Analyst]

@Giacomo Giuliani, yes though to not cover the line plots you might plot (or re-plot) the line plots after you've plotted the patches between the curves.

[Giacomo Giuliani]

It came to my minds one second after I posted. It works! Thanks.

Answer 4

[Adam Danz]

Another solution that may be helpful is the fillBetweenAreaCurve function offered by the MATLAB Charting team on the File Exchange.

[Les Beckham]

FYI - it appears that this File Exchange utility requires the Mapping Toolbox.