How to draw a histogram of my data?

I have two vectors a and b, where they are as follows:

b=1:100;

a=1e-18:1e-15;

I want to draw a histogram of them such that a is on x-axis and b is on y-axis as shown in the given attachement.

[Jon]

I think this was answered previously in https://www.mathworks.com/matlabcentral/answers/102406-how-can-i-plot-a-histogram-with-a-logarithmic-x-axis

[Sadiq Akbar]

Thank you very much Jon for your prompt response. I checked that but that has a single vector x. I have two vectors a and b.So If I make my "a" vector as x, then where is the role of vector b? So that can't help me.

Answer 1

[Jon]

I'm not really sure what you are trying to make a histogram of. It looks like you just want to know how many equally spaced values there are between logarithmically spaced intervals. In this case you don't need the b vector, other than to look at its maximum value which would just set the maximum horizontal scale value. To make sure you don't exceed this value you have to choose your increment properly. Maybe you are looking for something like this?

numBins = 20;
maxCount = 100;

% calculate increment so that maxCount will not be exceeded
delta = (1e-15 - 1e-18)/(numBins*maxCount);

% generate your equally space data
a=1e-18:delta:1e-15;

% make your histogram
histogram(a,numBins)
set(gca,'xscale','log') %scale the x-axis logarithmically

[Sadiq Akbar]

Thank you very much jon for your 2nd prompt response. If you look at the attachement, you will know what I want. As you can see in the attachment, (1) different bars have different coulurs, (2) There is a preview also, (3). Note the difference between the x-axes of both original and the preview. I want like that. I tried your 2nd try also, but it gives me (1) bars of all blue colurs,(2) There is no preview.

[Jon]

I see that you are trying to plot a histogram with different colors for each bar and a logarthimic x axis. I'm not quite sure how you are obtaining the bin counts to be plotted in your histogram. I'm not really seeing this in your data which you assign to the variable a, which seems to just be some uniformly spaced numbers. Putting that aside, and assuming you can figure out how to get the bin counts that you want plotted, here is an example of how you can plot them with similar colors to what you show. You can pick new colors using any RGB color combinations. See for example https://www.rapidtables.com/web/color/RGB_Color.html, and note that you have to divide these integer values, by 255 for MATLAB which wants them in the range 0-1.

numBins = 10;
colors = [76 0 153;0 102 204;0 153 255;255 153 51;255 255 0]/255 % define first 5 colors
numColors = size(colors,1);

% set x axis bin locations
x =  linspace(5e-18,5e-17,numBins)

% assign bin counts
% here I just hard code them for the example, but you should get them from
% a histogram of your data
N = [100,99,99,97,100,1,3,2 4 8]; 

% plot the histogram using a bar chart which allows control over colors etc
figure
h = bar(x,N,'FaceColor','flat','BarWidth',1);
h.CData = colors(mod(1:numBins,numColors)+1,:); % cycle through the colors

% use log scaling for x axis
set(gca,'xscale','log')

% plot again with expanded y axis
figure
h = bar(x,N,'FaceColor','flat','BarWidth',1)
ylim([96,101])
h.CData = colors(mod(1:numBins,numColors)+1,:); % cycle through the colors

% use log scaling for x axis
set(gca,'xscale','log')

[Sadiq Akbar]

Thank you very much Jon for your 3rd prompt response. Yes now it looks what is given in the attachment. But again there is a problem. If you see there is a PREVIEW also in the given attachment. How will we bring that preview here in this plot?

[Jon]

You can do it like this (note first plot is as before, second plot is now inset in first plot to provide "preview")

numBins = 10;
colors = [76 0 153;0 102 204;0 153 255;255 153 51;255 255 0]/255; % define first 5 colors
numColors = size(colors,1);

% set x axis bin locations
x =  linspace(5e-18,5e-17,numBins);

% assign bin counts
% here I just hard code them for the example, but you should get them from
% a histogram of your data
N = [100,99,99,97,100,1,3,2 4 8]; 

% plot the histogram using a bar chart which allows control over colors etc
figure
h = bar(x,N,'FaceColor','flat','BarWidth',1);
h.CData = colors(mod(1:numBins,numColors)+1,:); % cycle through the colors

% use log scaling for x axis
set(gca,'xscale','log')

% plot "preview" with expanded y axis inset within existing figure
% you can adjust postion values to get inset size and position just the way
% you want
axes('Position',[0.7 0.7 0.2 0.2])
box on
h = bar(x,N,'FaceColor','flat','BarWidth',1);
ylim([96,101])
h.CData = colors(mod(1:numBins,numColors)+1,:); % cycle through the colors

% use log scaling for x axis
set(gca,'xscale','log')

[Sadiq Akbar]

Thank you very much Jon for your prompt responses and help. Indeed it works. But there is a question. Why does it not work in Matlab2016b and it works in Matlab2020a?

In Matlab2016b, it gives the following error:

No public property CData exists for class matlab.graphics.chart.primitive.Bar.

Error in HistogrambyJon (line 13)

h.CData = colors(mod(1:numBins,numColors)+1,:); % cycle through the colors

I

[Jon]

If it will not run in the older version, then I assume that this feature was added in a more recent version. I don't have an older version to test in. Unless there is something preventing you from doing so I would highly recommend staying up to date with the current version to avoid such problems.

[Jon]

If I have answered your question please accept this answer so that others who might have a similar question will know that an answer is availble.