Question

Along with 3 numeric values the data contains categorical value (0 or 1) and would like to display the data using 3D scatter plot. I have tried to write a function to read the data from csv file and create a scatter plot the following way:

function f = ScatterPlotUsers(filename)

    data = ReadCSV(filename);
    x = data(:,2);
    y = data(:,3);
    z = data(:,4);

    size = ones(length(x), 1)*20;
    userType = data(:,5);
    colors = zeros(length(x), 1);

    a = find(userType == 1);
    b = find(userType == 0);
    colors(a,1) = 42; %# color for users type a
    colors(b,1) = 2;  %# color for users type b

    scatter3(x(:),y(:),z(:),size, colors,'filled') 
    view(-60,60);

What I actually wanted to do was to set the colour for a to red and b to blue but no matter the colour values (42 and 2 in example) the colours of dots don’t change.
Does anyone know what is the right way to determine specific colours for several categorical values (in this case only 0 and 1)?

Answer 1

You are doing it right, however, are you sure the colormap entries 42 and 2 refer to red and blue? You could try to give RGB values explicitly:

colors = zeros(length(x), 3);
colors(userType == 1, :) = [1 0 0]; %# colour for users type a (red)
colors(userType == 2, :) = [0 0 1]; %# colour for users type b (blue)

Also, I’d advise you to change the name of the variable size, since size is also a Matlab command; Matlab might be confused about that, and any future readers of your code will certainly be confused about that.