Label index/variable name near data points

Hello I,m looking for solution to print a index near to data point:

The example values are:

M =
        2278           2         685           2
         197           2        2221           2
         524           2        2496           2
        1061           2        3415           2
         651           2        5810           2
        1105           2        3817           2
         276           2        4527           2

I want to have a text box near to (2278,2) 1; (197,2) 2 .....

lables = 1 %lables = (1:k) 
figure('Name','Result','NumberTitle','on');
hold on
p11 = plot(M(:,2),M(:,1),'ro','Color','g');
hold on
p12 = plot(M(:,4),M(:,3),'rx','Color','r');
text(M(:,4),M(:,3),labels,'VerticalAlignment','bottom','HorizontalAlignment','right')
xlabel('tube dimension')
ylabel('Results')
tube_size_help = sprintf('%.0f',tube_size);
nameX=strcat(name ,'_tube_size_',tube_size_help);
saveas(gcf, [nameX '_results.png'])

But I get everythere the same value. Hov can I insert a changing index of row near to M(row,2),M(row,1) (or names of measures)

for loop? Or somethink better?

Thank you!

Answer 1

[Adam Danz]

This is easy with labelpoints() found on the file exchange.

figure('Name','Result','NumberTitle','on');
hold on
p11 = plot(M(:,2),M(:,1),'ro','Color','g');
hold on
p12 = plot(M(:,4),M(:,3),'rx','Color','r');

labelpoints(M(:,4),M(:,3), 1:size(M,1), 'E', 0.2)
% "E" means "east" of the coordinate
% 0.2 is a buffer space between the label and the marker

[Nik Rocky]

Genius! Thank you very much! :)

[Adam Danz]

Glad I could help!

[Nik Rocky]

Hello, do you have maybe idea, how to avoid overlapping of values? I'm allready turn TP to east and FP to ost but some values are still overlapping. Thank you!

[Adam Danz]

There is no automated way to detect and reposition overlapping labels but I have thought of this idea and may implement it in the future so be sure to "follow" that file exchange page so you get updates whenever I update the file.

Here's a workaround.

The 5th input below is the "buffer" value (default=0) which determines the space between the (x,y) coordinate and the label. The value is normalized to 10% of the axis limits so a value of 1 will space the label 10% of the axis limit from the label. The documentation describes 'buffer' as a scalar value but it can also be a vector as shown in the example under "idea 1", below.

labelpoints(x,y,'label','E',0.2);  % Example

Idea 1: you could shuffle the buffer values for all labels. This example below uses 3 different buffer values and cycles between those values.

x = ones(1,9); 
y = rand(1,9); 
subplot(3,1,[1,2])
plot(x,y,'o')
% Cycle through 3 buffers to get 9 values
buffer = [.1 .3 .5];
buffer = repmat(buffer,1,ceil(numel(y)/numel(buffer))); 
buffer(numel(y)+1:end) = []; 
% Sort the buffer values so they match the sorted y values
[~, ySortIdx] = sort(y); 
buffer(ySortIdx) = buffer; 
% Label points
labelpoints(x,y,1:numel(y),'E',buffer)

Idea 2: you could compute the distance between all pairs of points and somehow use that to determine the buffer value. I haven't thought it through the whole way so there might be some difficult hurdles but it seems doable.

The y-distance between all pairs of points can be computed using

dist = squareform(pdist(y')); 
dist(triu(true(size(dist)))) = nan; 

[Nik Rocky]

Thank you very much! I used your first idea - its looks good!

[Adam Danz]

Good! I should also mention that you can click-and-drag the labels after they are plotted, too, in case you need to make final adjustments to a figure you intend to present or publish.