Scatterhist function not allowing for boxplot on the y-axis to stay

Hi I'm trying to run a scatterhistogram (scatterhist()) but whenever I run the function the figure will pop up with the boxplot correctly displayed for a brief second on the y-axis and then disappear immediately. I'm not sure why. I also want to be able to color in the points and I'm not sure how to do that. Any help will be appreciated.

k = scatterhist(swine2, ready.rr_la, 'Group', swine);
hold on;
j = get(k(1),'colororder');
boxplot(k(3),ready.rr_la,swine2, 'Labels', {'10','11', '12', '13'}, 'color', j);
view(k(2),[270,90]);
xlabel('Subject');
ylabel('Err_la');

The figure displayed is atttached

[VBBV]

Try with view option using same handle as boxplot

 % if true
    % code
 % end
view(k(3),[270,90])
axis(k(1),'auto') % sync axes

Use label instead of labels

 % if true
    % code
%  end 
boxplot(k(3),ready.rr_la,swine2, 'label',...)

[Saar Peles]

This didn't work, it only pushed the plot itself to the right

[Adam Danz]

Please provide a minimal working example that demonstrates the problem.

When I add a boxplot to a scatterhis, it works fine.

[Saar Peles]

Sure, when adding all the view functions suggested this is what I get. When it first loads, the boxplot pops up but then immediatedly disappears.

[Adam Danz]

That's helpful but it's not a "MWE" (something we can copy-paste and run on our end to reproduce the problem).

It looks like you're not specifying the axes. I'll add an answer but if it doesn't solve the problem you'll need to give a MWE.

Answer 1

[Adam Danz]

It looks like you're trying to add a boxplot to the y-historgram axes. Recall that the y-histogram in scatterhist is reversed. Also, you're rotating the scatter axes so that x and y are switched but the histograms aren't switched. Why not just switch x and y in the scatterhist function?

h = scatterhist(y,x,___); 

To add a boxplot to the y-histogram axis,

hold(h(3), 'on')
boxplot(h(3), ___)

Use h(2) for the x-histogram or h(1) for the scatter axes.

Update: see this example in the Matlab documentation.

[Saar Peles]

Hi think I get what you're saying. I attempted the fix and this is what I got. I'll provide you with the MWE.

subject = ["S10", "S11", "S12", "S13"];

swine = categorical({'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S10', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S11', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S12', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13', 'S13'});

infAnt = categorical({'Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Aant', 'Septal', 'Ainf', 'Alat','Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Aant', 'Septal', 'Ainf', 'Alat','Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Aant', 'Septal', 'Ainf', 'Alat','Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Anterior', 'Anterior', 'Inferior', 'Inferior', 'Inferior', 'Anterior', 'Aant', 'Septal', 'Ainf', 'Alat'});

wallSept = categorical({'Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Wall', 'Wall','Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Wall', 'Wall','Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Wall', 'Wall','Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Septal', 'Wall', 'Wall', 'Wall', 'Wall', 'Septal', 'Wall', 'Wall'});

depth = categorical({'Base', 'Base', 'Base', 'Base', 'Base', 'Base', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Apex', 'Apex', 'Apex', 'Apex','Base', 'Base', 'Base', 'Base', 'Base', 'Base', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Apex', 'Apex', 'Apex', 'Apex','Base', 'Base', 'Base', 'Base', 'Base', 'Base', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Apex', 'Apex', 'Apex', 'Apex','Base', 'Base', 'Base', 'Base', 'Base', 'Base', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Mid', 'Apex', 'Apex', 'Apex', 'Apex'});


swine2 = [ 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13];
k = scatterhist(swine2, ready.rr_la, 'Group', swine);
hold(k(3), 'on');
j = get(k(1),'colororder');
boxplot(ready.rr_la,swine2,k(3), 'labels', {'10','11', '12', '13'}, 'color', j);
view(k(2),[270,90]);
xlabel('Subject');
ylabel('Err_la');

I also attached the struct that you'll need to load in order for this to work

[Adam Danz]

Thanks, that made things easy.

Problem #1: The axis handle needs to be the first input

% boxplot(ready.rr_la, swine2, k(3), 'labels', {'10','11', '12', '13'}, 'color', j);
%                 wrong spot ^^^^

Problem #2: The bigger problem; note the vertical axis limit in the originl data. I added axis(k(3),'on') to see th y-histogram axis ticks. Note that the vertical axis is the "X" axis and the horizontal axis is the y. It may be helpful to add ylabel(k(3),'y'); xlabel(k(3),'x'). Also note that the y-ticks in the histogram match the y-ticks in the scatter plot; same with the x-ticks for the lower histrogram axes. You'll want to maintain this relationship, otherwise the histograms are useless and misleading.

Now add the boxplots,

The origina data are still there off the axes. You can see them if you change the axis limits but now the axes are useless and no longer align with the scatter plot.

xlim(k(3),[-.2 5])
ylim(k(3),[-.2, 15])

So, what to do about this!? Well, I don't understand the point of the boxplots being there in the first place. If they relate to the histogram, you'll need to normalize them so they fit on top of the hist rather than somehwere else in the axes. Recall from the documentation, "By default, scatterhist finds the optimal bandwidth for estimating normal densities".

[Saar Peles]

I appreciate you taking the time to do this, but how would I put the boxplots onto the data. I tried doing it with the example you sent in my other post, however the erro it gave me was strange to say the least

[Adam Danz]

I don't understand how your boxplots are supposed to map onto the histograms. But their x & y values need to be within range of the histogram's x & y values.

That will be very hard to do.