How to assign error bar at middle of histogram?

Hello everyone,

I am trying to plot the histogram with the error bar at the middle point of histogram, and I used 'bar' syntax to get the plot, but couldn't get what I was looking for. Any help will be really appreciated.

% bin edges
bi = 0 :3: 175;%bin interval
% absolute frequencies
af1 = histc(VarName1, bi);
af2 = histc(VarName2, bi);
af3 = histc(VarName3, bi);
af4 = histc(VarName4, bi);
af5 = histc(VarName5, bi);
%VarName1, VarName2,VarName3,VarName4,VarName4 are my five distribution of data
mean=(af1+af2+af3+af4+af5)/5.0;
error=sqrt(((mean-af1).^2+(mean-af2).^2+(mean-af3).^2+(mean-af4).^2+(mean-af4).^2)/5);
%histogram(mean,be)
bar(bi,mean,'histc');
hold on
errorbar(bi,mean,error,'k','linestyle','none')

Thanks,

AG

Answer 1

[OCDER]

Here are some suggestions. To see what each line is doing, remove the ";".

%Store your data in cells so you can use loops. Example:
VarName = cell(1, 5);
for j = 1:5
    VarName{j} = randi(175, 100, 1);
end

%Label your variables with 1st-letter cap, to prevent overwriting matlab
%functions. For instance, you had "mean = (af1+af2+..." which would 
%override the "mean.m" matlab function! This leads to a lot of bugs.
Bin = 0:3:175;
AF  = zeros(length(Bin), length(VarName)); %Store your binned data in a matrix
for j = 1:5
    AF(:, j) = histc(VarName{j}, Bin);
end
Mean  = mean(AF, 2);

%error=sqrt(((mean-af1).^2+(mean-af2).^2+(mean-af3).^2+(mean-af4).^2+(mean-af4).^2)/5
%Notice you have a typo. (mean-af4).^2 comes twice. 
%To prevent coding error, use following practices:  
Error = sqrt(sum((Mean-AF).^2, 2)/5)
%Error = std(AF, [], 2); %Did you want standard dev?

%Plot data and store the handles as variables. Makes it easier to adjust.
Bx = bar(Bin, Mean, 'histc');
hold on 
Shift = diff(Bin(1:2))/2; %Need to shift your error bar to center it
Ex = errorbar(Bin+Shift, Mean, Error, 'k', 'linestyle', 'none');
hold off

%Adjust plots via those handles. Example:
Bx.FaceColor = [1 1 1]; %Make bars white
Ex.Color = [1 0 0]; %Make error bars red

[Ashok Gurung]

If I already have my variable on work space with 5 columns and unequal number of rows, then how should I proceed in first step? For example VarName = data_files(96,5). Can I proceed with line 9 in your code? Thanks for your help and I really appreciate that. AG

[OCDER]

What is the size of VarName1, VarName2, etc? If you're dealing with a 2D matrix each, then you'll have to modify the code to handle these. But then, are you trying to plot 5 distributions of 5 categories of data in a 3D histrogram?

OR, you could plot per each column of data in its own plot, which is simpler and can be done by wrapping the code above in a for loop:

%Plot data separately per each column of VarNameN
 for w = 1:5 
    VarName = {VarName1(:, w), VarName2(:, w), VarName3(:, w), VarName4(:, w), VarName5(:, w)};

    %COPY & PASTE CODE ABOVE, FROM LINE 9 TO END
 end

[Ashok Gurung]

the size of VarName1 is 96X1, VarName2 is 86X1, VarName3is 95X1 VarName4 is 90X1 and VarName5 is 98X1, how can I read each column from work space table?

[OCDER]

The short answer is to not store the variables as VarName1, VarName2, VarName3, VarName4, VarName5 in the first place. This is a common issue when using Matlab as discussed here:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

If you already have the variables in the workspace, store into a new cell array as such:

 VarName = {VarName1, VarName2, VarName3, VarName4, VarName5};

 %Then use the initial code above in the answer section, starting from

 Bin = 0:3:175;
 AF  = zeros(length(Bin), length(VarName)); %Store your binned data in a matrix
 for j = 1:5
    AF(:, j) = histc(VarName{j}, Bin);
 end
 Mean  = mean(AF, 2);
 ...THE REST OF THE CODE...

[Ashok Gurung]

Hey OCDER, I know it takes a long time to reply back, apology for that. I checked the code today, and it's working now. Thanks for the help and I really appreciate that. THX AG

[OCDER]

You're welcome!