Is there any way to add a set of random punctual z values labels inside the plot using a contourf x, y, z command?

Is there any way to automatically add a set of random punctual z values labels inside the plot using a contourf x, y, z command (in addition to contour lines labels shown by default)?

As reference see picture attached in which I manually added text box using Paint .

Answer 1

[Adam Danz]

Add a set of random z value labels inside a contourf plot

This demo shows the following.

1. Define x (vector), y (vector), and z (matrix) values for a contourf plot. (If x and y are matrices, see note below)
2. Randomly sample the x and y vectors.
3. Using the randomly sampled x and y values, pull out the corresponding z data to create the text labels.
4. Use text to add labels.

See code comments for details.

rng('default') % For consistent results

% Contour data
z = peaks();
x = 1:width(z); 
y = 1:height(z); 

% plot contour
contourf(x,y,z,"ShowText",true)

% Sample some points within the contour plot
n = 10;  % number of random labels

xRandIdx = randperm(numel(x),n);
yRandIdx = randperm(numel(y),n);
zIdx = sub2ind(size(z),yRandIdx,xRandIdx); 

% Create text labels
tx = x(xRandIdx); 
ty = y(yRandIdx); 
tz = compose('%.2f',z(zIdx)); % 2 decimal places
text(tx,ty,tz, ...
    'HorizontalAlignment','center',... % Center the label horizontally 
    'VerticalAlignment','middle',...  % Center the label vertically
    'FontSize', 8, ...
    'Color',[.4 .4 .4], ...
    'EdgeColor',[.4 .4 .4], ...  % Show the box frame
    'Margin', .2)  % make the box frame tighter

colorbar() % verify results by comparing values to colorbar

If x and y are matrices, generate text inputs like this instead of the above.

randIdx = randperm(numel(x),n);

% Create text labels
tx = x(randIdx); 
ty = y(randIdx); 
tz = compose('%.2f',z(randIdx)); % 2 decimal places

[MARCO BALESTRINI]

Hi Adam,

thank you very much! I think your answer it's the only one that almost does the trick.

The problem is now the script outputs just a little section of the original plot...I don't understand why.

This is my try:

[file, location] =uigetfile(' .xlsx');
addpath(location)
table = readtable(file);

Data=[table.n_Set, table.T_Set, table.T_GAS_TURB];

VariableName={'Speed','Torque','T_GAS_TURB'};
Datatable=array2table(Data,'VariableNames',VariableName);

ind_eff_neg=find(Datatable.T_GAS_TURB<0);
Datatable(ind_eff_neg,:)=[];

speed_val = unique(Datatable.Speed);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    length_speed(i) = length(index);
    clear index
end

clear i

[ MaxLenght,  ~] = max(length_speed);

x = NaN(length(speed_val),  MaxLenght);
y = NaN(length(speed_val),  MaxLenght);
z = NaN(length(speed_val),  MaxLenght);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    
    x(i,1:length(index)) = ones(1, length(index)).*speed_val(i);
    y(i,1:length(index)) = Datatable.Torque(index);
    z(i,1:length(index)) = Datatable.T_GAS_TURB(index);
end
 
rng('default') % For consistent results


z = peaks();
x = 1:width(z); 
y = 1:height(z); 

surface(x,y,z)

contourf(x,y,z,0:25:600,':', "ShowText",true,'LabelSpacing',800)


n = 10;  % number of random labels
xRandIdx = randperm(numel(x),n);
yRandIdx = randperm(numel(y),n);
zIdx = sub2ind(size(z),yRandIdx,xRandIdx); 

tx = x(xRandIdx); 
ty = y(yRandIdx); 
tz = compose('%.2f',z(zIdx)); % 2 decimal places
text(tx,ty,tz, ...
    'HorizontalAlignment','center',... % Center the label horizontally 
    'VerticalAlignment','middle',...  % Center the label vertically
    'FontSize', 8, ...
    'Color',[.4 .4 .4], ...
    'EdgeColor',[.4 .4 .4], ...  % Show the box frame
    'Margin', .2)  % make the box frame tighter

map = interp1([0; 0.35 ; 1], [0 1 0; 1 1 0; 1 0 0],linspace(0,1,100));

colormap(map)

grid on
colorbar eastoutside
clim([0 650])

title 'Temp after turbine[°C]'
xlabel 'Speed [RPM]'
ylabel 'Torque [Nm]'

ticks_x = 900:200:2300;
ticks_y = 100:200:2500;

xticks(ticks_x)
yticks(ticks_y)

[Adam Danz]

Be careful about blindly copy-pasting code. Let's take a look at what my solution is doing.

This line sets the random number generator so that each time you run it, it selects the same set of random text labels. You may not want to use this line .

rng('default') % For consistent results

This section generates contour data. You don't need this section because you have your own data to use.

% Contour data
z = peaks();
x = 1:width(z); 
y = 1:height(z); 

This line creates the contourf plot. Modify it to do whatever you were doing in your original work.

% plot contour
contourf(x,y,z,"ShowText",true)

This is main focus to learn from. n is the number of text lables to randomly generate. It looks like you are also using x y z variable names. In my demo, x and y are vectors. z must be a matrix. If your x and y variables are matrices, you need to change this a bit (see the bottom of my answer).

% Sample some points within the contour plot
n = 10;  % number of random labels

xRandIdx = randperm(numel(x),n);
yRandIdx = randperm(numel(y),n);
zIdx = sub2ind(size(z),yRandIdx,xRandIdx); 

This section creates the text labels and adds a colorbar that you can reference to check that the labels make sense.

% Create text labels
tx = x(xRandIdx); 
ty = y(yRandIdx); 
tz = compose('%.2f',z(zIdx)); % 2 decimal places
text(tx,ty,tz, ...
    'HorizontalAlignment','center',... % Center the label horizontally 
    'VerticalAlignment','middle',...  % Center the label vertically
    'FontSize', 8, ...
    'Color',[.4 .4 .4], ...
    'EdgeColor',[.4 .4 .4], ...  % Show the box frame
    'Margin', .2)  % make the box frame tighter

colorbar() % verify results by comparing values to colorbar

[MARCO BALESTRINI]

Hi Adam, I'm sorry...I'm a total beginner in coding :-(

My x, y and z are Tables taken form a .xlsx file single columns, so plain vectors .

This is my try:

[file, location] =uigetfile(' .xlsx');


addpath(location)
table = readtable(file);
Data=[table.n_Set, table.T_Set, table.T_GAS_TURB];


VariableName={'Speed','Torque','T_GAS_TURB'};
Datatable=array2table(Data,'VariableNames',VariableName);

ind_eff_neg=find(Datatable.T_GAS_TURB<0);
Datatable(ind_eff_neg,:)=[];

speed_val = unique(Datatable.Speed);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    length_speed(i) = length(index);
    clear index
end

clear i

[ MaxLenght,  ~] = max(length_speed);

x = NaN(length(speed_val),  MaxLenght);
y = NaN(length(speed_val),  MaxLenght);
z = NaN(length(speed_val),  MaxLenght);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    
    x(i,1:length(index)) = ones(1, length(index)).*speed_val(i);
    y(i,1:length(index)) = Datatable.Torque(index);
    z(i,1:length(index)) = Datatable.T_GAS_TURB(index);
end
 
surface(x,y,z)

contourf(x,y,z,0:25:600,':', "ShowText",true,'LabelSpacing',800)

rng('default')

n = 10;
xRandIdx = randperm(numel(table.n_Set),n);
yRandIdx = randperm(numel(table.T_Set),n);
zIdx = sub2ind(size(table.T_GAS_TURB),yRandIdx,xRandIdx);


map = interp1([0; 0.35 ; 1], [0 1 0; 1 1 0; 1 0 0],linspace(0,1,100));

colormap(map)

grid on
colorbar eastoutside
clim([0 650])

title 'Temp after turbine[°C]'
xlabel 'Speed [RPM]'
ylabel 'Torque [Nm]'

ticks_x = 900:200:2300;
ticks_y = 100:200:2500;

xticks(ticks_x)
yticks(ticks_y)

tx = x(xRandIdx); 
ty = y(yRandIdx); 
tz = compose('%.2f',z(zRandIdz)); 
text(tx,ty,tz, ...
    'HorizontalAlignment','center',...
    'VerticalAlignment','middle',...  
    'FontSize', 10, ...
    'Color',[.2 .2 .2], ...
    'EdgeColor',[.2 .2 .2], ...  
    'Margin', .5)  

colorbar()

but I still got an error using sub2ind: "Out of range subscript"

What are the subscripts that the "sub2ind" command needs to convert?

[Adam Danz]

I assume this line below, copied from your code, produces the expected contour plot.

contourf(x,y,z,0:25:600,':', "ShowText",true,'LabelSpacing',800)

Looking at your code, x and y appear to be matrices so you would apply the lines of code at the bottom of my answer that starts with "If x and y are matrices...".

contourf(x,y,z,0:25:600,':', "ShowText",true,'LabelSpacing',800)

rng('default')

n = 10;

% REMOVE THESE LINES
% xRandIdx = randperm(numel(table.n_Set),n);
% yRandIdx = randperm(numel(table.T_Set),n);
% zIdx = sub2ind(size(table.T_GAS_TURB),yRandIdx,xRandIdx);


randIdx = randperm(numel(x),n);
% Create text labels
tx = x(randIdx); 
ty = y(randIdx); 
tz = compose('%.2f',z(randIdx)); % 2 decimal places

text(tx,ty,tz, ...
    'HorizontalAlignment','center',...
    'VerticalAlignment','middle',...  
    'FontSize', 10, ...
    'Color',[.2 .2 .2], ...
    'EdgeColor',[.2 .2 .2], ...  
    'Margin', .5)  

[MARCO BALESTRINI]

It works great!!!!!!

Thank you very very very very much!!!

(Sorry I still have to understand the difference between arrays and matrices in Matlab :-( ....)

But why the script shows me only 3 random points if I have n=10? ( if I increase the "n" value I get more points, but still less than "n").

Last question...if instead of random points I wanted the Z values in every grid node (points of x,y coordinates from my input data table) how could I edit my script?

[Walter Roberson]

(Sorry I still have to understand the difference between arrays and matrices in Matlab :-( ....)

In MATLAB:

• an empty array is typically 0 by 0, but can be any size provided that one or more of the dimensions is 0. There are cases where empty arrays will not be recognized as being empty unless it is 0 x something or something by 0.
• a scalar is 1 x 1 only
• a vector is 1 x something or something x 1. vectors include scalars, 1 x 1
• a matrix is something by a possibly different something. matrices include vectors (and scalars)
• an array is typically 3 or more dimensions. Arrays include matrices (and vectors and scalars)

[MARCO BALESTRINI]

Thank u very much Walter!

[Adam Danz]

If you're ever unsure how MATLAB defines these terms, see ismatrix, isvector, isscalar.

> why the script shows me only 3 random points if I have n=10?

I need more info to answer that. One possibility is that you are zooming into a section of your contour by setting the axis limits, although I'd expect to see the occasional text object appear outside of the axes since the Clipping property of text is off by default.

If you provide an updated version of your code I might get a better sense of what's going on.

> .if instead of random points I wanted the Z values in every grid node....

That sounds like it will be a lot of text but here's how:

Instead of this,

randIdx = randperm(numel(x),n);
% Create text labels
tx = x(randIdx); 
ty = y(randIdx); 
tz = compose('%.2f',z(randIdx)); % 2 decimal places

replace the data with your actual contour data

% Create text labels
tx = x(:); 
ty = y(:); 
tz = compose('%.2f',z(:)); % 2 decimal places

[MARCO BALESTRINI]

Adam I don't know how to thank you for your prompt help ❤ May I offer you a "virtual" coffe via Paypal (if it's allowed by the forum regulation)?

I confirm you that the "random" version shows only the 30% of the set "n" (that's 3 if n=10, 6 if n=20 etc etc)...but it's not a big deal...I can simply increase "n"

clc
[file, location] =uigetfile(' .xlsx');


addpath(location)
table = readtable(file);
Data=[table.n_Set, table.T_Set, table.T_GAS_TURB];


VariableName={'Speed','Torque','T_GAS_TURB'};
Datatable=array2table(Data,'VariableNames',VariableName);

ind_eff_neg=find(Datatable.T_GAS_TURB<0);
Datatable(ind_eff_neg,:)=[];

speed_val = unique(Datatable.Speed);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    length_speed(i) = length(index);
    clear index
end

clear i

[ MaxLenght,  ~] = max(length_speed);

x = NaN(length(speed_val),  MaxLenght);
y = NaN(length(speed_val),  MaxLenght);
z = NaN(length(speed_val),  MaxLenght);

for i = 1 : length(speed_val)
    index = find(Datatable.Speed==speed_val(i));
    
    x(i,1:length(index)) = ones(1, length(index)).*speed_val(i);
    y(i,1:length(index)) = Datatable.Torque(index);
    z(i,1:length(index)) = Datatable.T_GAS_TURB(index);
end
 
surface(x,y,z)

contourf(x,y,z,0:25:600,':', "ShowText",true,'LabelSpacing',800)

rng('default')

n = 10;

randIdx = randperm(numel(x),n);

tx = x(randIdx); 
ty = y(randIdx); 
tz = compose('%.0f',z(randIdx));
text(tx,ty,tz, 'HorizontalAlignment','center', 'VerticalAlignment','middle','FontSize', 10,'Color',[.2 .2 .2],'EdgeColor',[.2 .2 .2],'Margin', .5)  

colorbar()

map = interp1([0; 0.35 ; 1], [0 1 0; 1 1 0; 1 0 0],linspace(0,1,100));

colormap(map)

grid on
colorbar eastoutside
clim([0 650])

title 'Temp after turbine[°C]'
xlabel 'Speed [RPM]'
ylabel 'Torque [Nm]'

ticks_x = 900:200:2300;
ticks_y = 100:200:2500;

xticks(ticks_x)
yticks(ticks_y)

[Adam Danz]

@MARCO BALESTRINI, I'm glad that worked out.

I noticed that your x, y, z matrices are preallocated as NaNs. The solution I proposed selects random values within the x,y,z data. If it selects locations and x and y that have NaN values, the text label will not appear. My hunch is that your data contain ~2/3 NaN values which is why only ~1/3 of the labes are appearing.

Thanks for the coffee offer! If you ever come to MathWorks HQ, let me know. We can meet for coffee and discuss turbine effects on temperature distributions 🤓.

Answer 2

[Taylor]

You can use the text command

[MARCO BALESTRINI]

Hi Taylor,

the problem is I want Matlab to automatically add some random punctual Z values between contour lines

[Taylor]

There are a variety of ways to generate random numbers in MATLAB

[MARCO BALESTRINI]

Maybe I miscommunicate the concept: I'd like my script could add a label containing the actual value of the Z variable in some points of the contourf x,y,z plot generated by the script, as in the pic attached.

Is it feasible?

[Walter Roberson]

@MARCO BALESTRINI

Selected_X_Points = rand(1,NumberOfPoints) * (max(X_Points)-min(X_Points)) + min(X_Points);

Repeat for Y.

Then

Selected_Z_Points = interp2(X_Points, Y_Points, Z_Points, Selected_X_Points, Selected_Y_Points);

You now have random x and y coordinates and corresponding Z coordinates for the purpose of text()

text(Selected_X_Points, Selected_Y_Points, string(round(Selectted_Z_Points)))

I somehow doubt that you actuall want random points, however.

[MARCO BALESTRINI]

Yes, I do want random points, because it's all aimed to help the plot readibility.

Anyway I can't understand your hint:

(max(X_Points)-min(X_Points)) + min(X_Points)   

Answer 3

[Umar]

Hi @MARCO BALESTRINI ,

To achieve your goal of adding labels that display the actual values of the Z variable at certain points on a contourf plot in MATLAB, you can use the ShowText option in combination with the LabelFormat property. Here is a step-by-step approach, start by defining your X, Y, and Z matrices. For example, using a mathematical function like peaks can serve as a good starting point:

   [X, Y, Z] = peaks(50);  % Create sample data

Use the contourf function to create the filled contour plot. You can specify levels if you want to control where the labels appear. For instance, to display contours at specified levels while showing text labels:

   contourf(X, Y, Z, 'ShowText', 'on');

If you want to format the labels (for example, to show one decimal place), you can use the LabelFormat property. As of R2022b, this allows you to define how labels are displayed:

   contourf(X, Y, Z, 'ShowText', true, 'LabelFormat', '%0.1f');

Here is complete example that combines all these steps:

   % Create sample data
   [X, Y, Z] = peaks(50);

   % Create filled contour plot with labeled contours
   contourf(X, Y, Z, 'ShowText', 'on', 'LabelFormat', '%0.1f');
   colorbar;  % Optional: adds a color bar for reference
   title('Filled Contour Plot with Labels');
   xlabel('X-axis');
   ylabel('Y-axis');

Now, if you are interested in showing labels in different units or applying custom formatting (like converting meters to feet), you can define a custom labeling function and pass it to LabelFormat. Here is an example function:

   function labels = mylabelfun(vals)
       feetPerMeter = 3.28084;
       feet = round(vals * feetPerMeter);
       labels = vals + " m (" + feet + " ft)";
       labels(vals == 0) = "0 m";
   end

   % Usage in contourf
   contourf(X, Y, Z, 'ShowText', true, 'LabelFormat', @mylabelfun);

Please see attached.

You can further enhance your plot's aesthetics by adjusting properties such as color maps (`colormap`), transparency (`FaceAlpha`), and line widths (`LineWidth`).

Hope this helps.

Please let me know if you have any further questions.