Is it possible to visualize all data in a multiple scatterplot?

Hi,

I am trying to make a scatterplot with a lots of data via this code:

h1 = scatter(table{:,variable},table{:,'PET'}, '+','MarkerEdgeColor',[102/255, 178/255, 255/255]);
hold on
h2 = scatter(table{:,variable},table{:,'SET_'}, '+','MarkerEdgeColor',[102/255, 255/255, 102/255]);
h3 = scatter(table{:,variable},table{:,'UTCI'}, '+','MarkerEdgeColor',[255/255, 204/255, 153/255]);
h4 = scatter(table{:,variable},table{:,'PT'}, '+','MarkerEdgeColor',[255/255, 102/255, 102/255]);
h5 = scatter(table{:,variable},table{:,'mPET'}, '+','MarkerEdgeColor',[204/255, 153/255, 255/255]); 

and the result is this:

It is obvious that the last plotted violet data will be above the others hiding the other colors. Is it possible to visualize all data to the same extend? For example it can be plotted through a rotation of all colours to achieve the desired visibility of all the data.

[Walter Roberson]

You can set Alpha properties for scatter() plots.

[Stepan Subik]

Hi, thank you for your suggestion. Unfortunatelly, this does not work in my case since I have too many data and the plot does not change with varying transparency.

Answer 1

[Ameer Hamza]

Maybe use scatter3() and visualize by using a z-offset. Not sure if it helps in your case

h1 = scatter3(table{:,variable},table{:,'PET'}, 0*ones(size(table{:,'PET'})), '+','MarkerEdgeColor',[102/255, 178/255, 255/255]);
hold on
h2 = scatter3(table{:,variable},table{:,'SET_'}, 1*ones(size(table{:,'SET_'})), '+','MarkerEdgeColor',[102/255, 255/255, 102/255]);
h3 = scatter3(table{:,variable},table{:,'UTCI'}, 2*ones(size(table{:,'UTCI'})), '+','MarkerEdgeColor',[255/255, 204/255, 153/255]);
h4 = scatter3(table{:,variable},table{:,'PT'}, 3*ones(size(table{:,'PT'})), '+','MarkerEdgeColor',[255/255, 102/255, 102/255]);
h5 = scatter3(table{:,variable},table{:,'mPET'}, 4*ones(size(table{:,'mPET'})), '+','MarkerEdgeColor',[204/255, 153/255, 255/255]); 

[Stepan Subik]

Hi, this is a great idea. But it is than less intuitive in the meaning of a comparison of those variables. I am searching for some way to plot it not a variable after variable, but points from every variable at once going from first variable to last n-times, where n is the number of points for one variable. Does something like this exist?

[Ameer Hamza]

Yes, it is possible. First you will need to create a column vectors for all the datapoints. Something like this

x = [table{:,variable}; table{:,variable}; table{:,variable}; table{:,variable}; table{:,variable}];
y = [table{:,'PET'}; table{:,'SET_'}; table{:,'UTCI'}; table{:,'PT'}; table{:,'mPET'}];
c = [0*ones(size(table{:,'PET'})); 1*ones(size(table{:,'SET_'})); 2*ones(size(table{:,'UTCI'})); 3*ones(size(table{:,'PT'})); 4*ones(size(table{:,'mPET'}))];

scatter(x, y, [], c);

you can try if this can improve the visualization quality.

[Stepan Subik]

Nice, now I have some long running program, but I will try that tomorrow ;)

[Stepan Subik]

Yes, it actually helped, thank you

[Ameer Hamza]

I am glad to be of help! Can you show the final figure and how it improved the visualization of all the points?