Adding a scatter of points to a boxplot

Does anyone come with with a code which can match the python generated boxplot?

[dpb]

Presuming the points actually a set of coordinates as shown, don't see why

hold on
scatter(x,y)

with appropriate x,y arrays and the associated color arrays, etc., wouldn't come reasonably close...

Answer 1

[dpb]

Indeed, that seems to work just fine...

load carsmall MPG              % the sample dataset variable
hold on
scatter(ones(size(MPG)).*(1+(rand(size(MPG))-0.5)/10),MPG,'r','filled')

yields

It's possible to add color with value scaling in scatter see the details on it for all the particulars.

[Peyman Obeidy]

Here is the same thing with a bit more manipulations

load carsmall MPG              % the sample dataset variable
MPG(:,2)=MPG(:,1).*2;
MPG(:,3)=MPG(:,1).*3;
boxplot(MPG,'Notch','on','Labels',{'mu = 5','mu = 6','mu = 6'},'Whisker',1)
lines = findobj(gcf, 'type', 'line', 'Tag', 'Median');
set(lines, 'Color', 'g');
% Change the boxplot color from blue to green

a = get(get(gca,'children'),'children');   % Get the handles of all the objects

%t = get(a,'tag');   % List the names of all the objects 

%box1 = a(7);   % The 7th object is the first box

set(a, 'Color', 'r');   % Set the color of the first box to green

hold on
x=ones(length(MPG)).*(1+(rand(length(MPG))-0.5)/5);
x1=ones(length(MPG)).*(1+(rand(length(MPG))-0.5)/10);
x2=ones(length(MPG)).*(1+(rand(length(MPG))-0.5)/15);
f1=scatter(x(:,1),MPG(:,1),'k','filled');f1.MarkerFaceAlpha = 0.4;hold on 
f2=scatter(x1(:,2).*2,MPG(:,2),'k','filled');f2.MarkerFaceAlpha = f1.MarkerFaceAlpha;hold on
f3=scatter(x2(:,3).*3,MPG(:,3),'k','filled');f3.MarkerFaceAlpha = f1.MarkerFaceAlpha;hold on

[dpb]

Nice choices..looks good. :)

[Peyman Obeidy]

Thank you

[Rubina Chandnani]

In the line of code in scatter, is there a way to use a different color using uisetcolor? (I don't want to use the default colors).

[Seth DeLand]

I'd like to add that there is now an easier way to do this with boxchart (added in R2020a) and swarmchart (added in R2020b):

load carsmall MPG   % the sample dataset variable
MPG(:,2)=MPG(:,1).*2;
MPG(:,3)=MPG(:,1).*3;
boxchart(MPG)
hold on
x = repmat(1:3,100,1);  % create the x data needed to overlay the swarmchart on the boxchart
swarmchart(x,MPG,[],'red')

Answer 2

[Hassan]

Hi, here you can find a one line solution for the jitter like function using the 'undocumented matlab' features.

load carsmall MPG 
figure; 
MPG(:,2)=MPG(:,1).*2;
MPG(:,3)=MPG(:,1).*3;
boxplot(MPG); 
hold on;
x=repmat(1:3,length(MPG),1);
scatter(x(:),MPG(:),'filled','MarkerFaceAlpha',0.6','jitter','on','jitterAmount',0.15);

Best, HM

[Junru Ruan]

This is the best answer! very nice presentation.

Tips: if you used group in box plot, do a 'unique' to get the right x axis.

    boxplot(report_table.data,report_table.group_id);
    hold on
    [C, ~, ic]= unique([report_table.group_id],'stable');
    scatter(ic,report_table.data,'filled','MarkerFaceAlpha',0.6','jitter','on','jitterAmount',0.15);
    xlabel('Group ID');
    ylabel('Data');
    hold off

Answer 3

[Ernesto Salcedo]

Boxchart solution for grouped categorical data

Table with random group

count = 20;
T = table(randi(10,count,1), categorical(repmat(["papaya";"silicon"], count/2,1)),'VariableNames',["Recharges","model"])
T.idx = grp2idx(T.model); % convert categories into group indices

Boxchart

figure
hc = boxchart(T.idx, T.Recharges); % group by index
hold on
% overlay the scatter plots
for n=1:max(unique(T.idx))
    hs = scatter(ones(sum(T.idx==n),1) + n-1, T.Recharges(T.idx == n),"filled",'jitter','on','JitterAmount',0.1);
    hs.MarkerFaceAlpha = 0.5;
end
set(gca,"XTick", unique(T.idx),"XTickLabel",categories(T.model))