How to update large data plot (collected over time), with multiple data series in MATLAP App?

Hello,

So I have an app, where I am collecting external data from 16 different sources. This data is constantly being updated, and I would like to have some sort of plot to show this data over time.

In essence, I have 16 different data sets, each with a value, and a timestamp.I would like these data sets to be plotted as different data series. As the data is collected, I update my UI. Currently, when I am updating my UI, I am replotting the entire data set for each data set. This gets very slow at even small numbers of data (i.e. over 5 mintues of data collection, I need this to run for a week).

This is what I want my graph output to look like (here I only have 3 data series, for debugging and testing purposes).

This is how I am updating my UI (everytime a data value is collected):

...%Pseudo code
    data = get_CV()
    update_main_graph(measurement_graph,data)
...
    
%Update UI function
function update_main_graph(measurement_graph,data)

for ii = 1:numel(fieldnames(data)) %number of fields is the number of data sources = 16
%Plot each measurement on the graph
    field_name = strcat("ring", int2str(ii));
    
    if isempty(data.(field_name).timestamp) ~= 1
    time_axis = datetime(data.(field_name).timestamp,'InputFormat','dd-MMM-yyyy HH:mm:ss');
    data_axis = data.(field_name).analyzed;
    
    scatter(measurement_graph, time_axis, data_axis);

    hold(measurement_graph,'on')
    
    pause(0.001)
    
    end
end

hold(measurement_graph,'off')

end

As you can see, I am replotting ALL THE DATA each time.

I can't figure out a way just to plot the new collected data every time.

Does anyone have an idea how I can go about doing this? I would really like to have a visualization of the data while the data is collecting, but if this won't be feasible I'll make do.

Any advice is GREATLY appreciated!!

Answer 1

[Mario Malic]

Can you check if time_axis and data_axis contain all the data from the measurements? If yes, you can try this

 scatter(measurement_graph, time_axis(end), data_axis(end)); % plots the last element

The issue might be with uifigure, since it can get slow. Try creating a figure() with axes() and do plotting on it, unfortunately it will be in a window outside the UIFigure from AppDesigner. Also 1 ms pause might be problematic with refreshing the window You can try as well

drawnow limitrate

[Mark Lepage]

The problem with this approach is that currently, inside the for loop - the data is plotted in a different data series everytime the scatter() function is called. This is what I want, as I want each data set plotted in a different series.

However, with your method, each time the new last element is plotted, it creates a new data series...

What I need is a method to call on a specific data series and to continue to plot just on that data series, i.e. something like

for ii = 1:numel(fieldnames(data))
%Plot each measurement on the graph
    field_name = strcat("ring", int2str(ii));
    
    if isempty(data.(field_name).timestamp) ~= 1
    time_axis = datetime(data.(field_name).timestamp,'InputFormat','dd-MMM-yyyy HH:mm:ss');
    data_axis = data.(field_name).analyzed;
    
    %% Here I call upon the data series, to continue plotting to it - which doesn't exist to my knowledge
    scatter(measurement_graph, field_name_data_series, time_axis, data_axis);
    hold(measurement_graph,'on')
end

[Mario Malic]

In scatter property list, there's XDataSource, XData, Y... and so on, maybe those properties would be of interest, combined with refreshdata.