2

I have a figure with two or more lines. These lines have additional, important information associated with them, like how many data points were averaged to create the line etc. I would like to access this information in my figure.

I thought a good solution for this would be to if you could hover over a line with your mouse and get that extended information.

However searching for tooltips/overlays/hover-over on figures seemed to not be fruitful.

Example:

figure; hold on;
plot(1:10,rand(10,1))
plot(1:10,rand(10,1))
% additional info
plot_1_info.name = 'Alice';
plot_2_info.name = 'Bob';
plot_1_info.age = 24;
plot_2_info.age = 12;

Any good solutions or better approaches for this?

Wolfie
  • 27,562
  • 7
  • 28
  • 55
user2305193
  • 2,079
  • 18
  • 39
  • 2 questions: 1) Which MATLAB version are you using? 2) Are UIFigures an option for you? – Dev-iL Mar 26 '19 at 11:30
  • 1) R2018b Update3, newest stable if I'm correct 2) I don't know about them I think, but I'd prefer as-simple-as-possible-approach – user2305193 Mar 26 '19 at 11:37
  • 1
    1) Newest stable is R2019a ;) 2) It's just as simple... more a question of compatibility with existing code/figures (uifigures are the "new" figures). – Dev-iL Mar 26 '19 at 11:42
  • both answers are awesome, chose Wolfie's one for backcompatibility, both should be right – user2305193 Mar 26 '19 at 14:35

2 Answers2

6

You can change the data cursor behaviour, this option has good backwards compatability (I've tested the below in R2017b, used similar before in 15b).

data cursor

See my comments for details:

% Create some data
x = (1:2:20).';
y = rand(10,1);
name = { 'Alice'; 'Alice'; 'Alice'; 'Alice'; 'Bob'; 'Bob'; 'Bob'; 'Chris'; 'Chris'; 'Chris' };
age = [ 24; 24; 24; 24; 12; 12; 12; 17; 17; 17 ];
% Put it in a table, so we have it all together for indexing as plot data
tbl = table( x, y, name, age );

% Create the plot, assign the UserData property to the plot object
f = figure; 
plt = plot( x, y );
plt.UserData = tbl;

% Hijack the Data Cursor update callback so we can inject our own info
dcm = datacursormode( f );
set( dcm, 'UpdateFcn', @onDataCursor );

% Function which returns the text to be displayed on the data cursor
function txt = onDataCursor( ~, evt )
    % Get containing figure
    f = ancestor( evt.Target, 'figure' );
    % Get the index within the original data
    idx = getfield( getCursorInfo( datacursormode( f ) ), 'DataIndex' );
    % The original data is stored in the UserData property
    data = evt.Target.UserData;
    % Each element of the cell array is a new line on the cursor
    txt = { sprintf( 'X: %g', data.x(idx) ), ...
            sprintf( 'Y: %g', data.y(idx) ), ...
            sprintf( 'Name: %s', data.name{idx} ), ...
            sprintf( 'Age: %g', data.age(idx) ) };          
end

Output:

plot with data tip

Note: I've not handled anywhere the case that there is more than one data cursor tip. You can easily implement a loop over idx within the callback to handle this, I leave it as an exercise.

This approach is really flexible. For instance if we had 3 lines (one per 'person') then they can each have their own UserData struct, and we don't need to repeat all the info in table rows.

A = struct( 'X', 1:4, 'Y', rand(1,4), 'Name', 'Alice', 'Age', 24 );
B = struct( 'X', 1:3, 'Y', rand(1,3), 'Name', 'Bob', 'Age', 12 );
C = struct( 'X', 1:3, 'Y', rand(1,3), 'Name', 'Chris', 'Age', 17 );

f = figure; hold on;
plt = plot( A.X, A.Y ); plt.UserData = A;
plt = plot( B.X, B.Y ); plt.UserData = B;
plt = plot( C.X, C.Y ); plt.UserData = C;

% ... Now the struct fields can be accessed from the callback
Wolfie
  • 27,562
  • 7
  • 28
  • 55
  • 1
    It's worth mentioning that [in R2019a customizing datatips was greatly simplified](https://www.mathworks.com/help/matlab/release-notes.html?rntext=&startrelease=R2019a&endrelease=R2019a&groupby=release&sortby=descending#mw_f3339c68-a89c-4961-8a50-b16957920f31). – Dev-iL Mar 26 '19 at 11:49
  • 1
    @Dev-iL Feel free to post an answer, I've only got up to R2017b for testing, and the OP doesn't have R2019a. – Wolfie Mar 26 '19 at 11:50
  • thanks for this answer I will try it. @Dev-iL: I'm upgrading right now, please do post your answe :) – user2305193 Mar 26 '19 at 12:19
5

Using the new data tip customization system introduced in R2019a, we can do the following:

figure(); hP = plot(1:10,rand(10,1),1:10,rand(10,1));
nPts = cellfun(@numel, {hP.XData});
hP(1).DataTipTemplate.DataTipRows(end+1) = dataTipTextRow('Name', repmat("Alice",nPts(1),1) );
hP(1).DataTipTemplate.DataTipRows(end+1) = dataTipTextRow('Age',  repmat(12,     nPts(1),1) );
hP(2).DataTipTemplate.DataTipRows(end+1) = dataTipTextRow('Name', repmat("Bob",  nPts(2),1) );
hP(2).DataTipTemplate.DataTipRows(end+1) = dataTipTextRow('Age',  repmat(24,     nPts(2),1) );
% (Of course the above can be organized in a nicer way using a function)

Which yields:

enter image description here

Notice that the "hover datatip" has black text, while the "click datatip" has blue text - this is default behavior.

Dev-iL
  • 23,742
  • 7
  • 57
  • 99