Histogram for Grouped Data

Good evening,

I'm wondering if there is a way to plot different histograms on the same graph exploiting grouping variables. I know the function boxplot does it on its own, but I'm, wondering if there is a way to do it using histogram(.).

Basically I want to plot the variation for consumption for different age grups/income quintiles/etc which I have stored as categorical variables.

Can anyone help?

Thank you very much in advance!

Lorenzo

[dpb]

W R2014b histcounts and histogram support categorical variables to at least some degree...w/o knowing precisely what your data are and what you want, whether directly out of the box solution or not, can't tell...

Answer 1

[the cyclist]

I don't believe there is a built-in grouping variable capability in histogram, but it is pretty easy to code

% Made-up data
N = 1000;
x = randn(N,1);

% Grouping variable will be either 0 or 1
gv = round(rand(N,1));

% Offset the gv=1 data, so the histograms are more distinct
x(gv==1) = x(gv==1) + 3;

% Plot the two histograms
figure
hold on
histogram(x(gv==0))
histogram(x(gv==1))
legend({'gv=0','gv=1'})

[Lorenzo Cito]

Thank you very much, I had tought of something similar. Do you think is it possible to make it work with "gv" being a cell array cointaining words rather than 1 and 0s?

Probably I'm coding something wrong as I don't think histogram(x(gv=='Old')) - for example - is the proper way to write it.

Thank you for your patience, I'm quite new with MatLab and programming in general!

[the cyclist]

There are many possible ways to do get the logical array you need from the cell. Here is one:

C = {'Old','New','New'}; % Your grouping
ismember(C,'Old'); % Logical variable pointing to only first one

[Lorenzo Cito]

Thank you very much!