2D Plot of CSV data

I would like to create a 2D plot with a colored surface (i.e. a xy coordinate system where the color indicates a temperature at each point). I did this before by using the contourf function, however, this function requires three matrices with the x, y, and color data of each point as entry.

The problem here: my data comes from a CSV file, where each line has three entries: x coordinate, y coordinate and temperature. Example:

X                        Y                        T (K)
31.999999999999964       0.19038773875061182      323.5661344306932
31.980976178007637       0.20000000000002974      323.58694805436954
31.999999999999957       0.20000000000003018      323.58694573460093

Unfortunately these coordinates don't follow a regular pattern and there is no regular grid I could generate from this data. So how can I generate a color plot from this?

(Regardless of my current problem: why does the countourf function require X and Y matrices? I found this annoying several times in the past when I had data given as a CSV with coordinates and data in one line, but usually my data had a regular grid as a basis and I could just generate the X and Y matrices using meshgrid(). However, even if I had a regular spacing for my x and y coordinates, I would still have to dig into my data to determine the periodicy of the grid first and as a bonus I will get various occasions to mess up x/y directions, table indices, etc. until I finally get it done correctly.)

Answer 1

[Voss]

One way you can do it (still with contourf()) is to use a scatteredInterpolant to interpolate and/or extrapolate T from the X,Y points where T is defined to all X,Y points in a grid:

X = [32; 31.98; 32];
Y = [0.19; 0.2; 0.2];
T = [323.566; 323.587; 323.587];
I = scatteredInterpolant(X,Y,T)

I = 
  scatteredInterpolant with properties:

                 Points: [3×2 double]
                 Values: [3×1 double]
                 Method: 'linear'
    ExtrapolationMethod: 'linear'

xc = 30:0.1:34;
yc = 0.17:0.01:0.22;
[XX,YY] = meshgrid(xc,yc);

figure();
contourf(XX,YY,I(XX,YY));
colorbar();

Another example with more points (random data this time):

X = 31+2*rand(50,1)-1;
Y = 0.2+0.2*rand(50,1)-0.1;
T = 323.5+2*rand(50,1)-1;
I = scatteredInterpolant(X,Y,T)

I = 
  scatteredInterpolant with properties:

                 Points: [50×2 double]
                 Values: [50×1 double]
                 Method: 'linear'
    ExtrapolationMethod: 'linear'

xc = 30:0.1:32;
yc = 0.1:0.01:0.3;
[XX,YY] = meshgrid(xc,yc);

figure();
contourf(XX,YY,I(XX,YY));
colorbar();

Answer 2

[Star Strider]

One approach —

M = [31.999999999999964       0.19038773875061182      323.5661344306932
    31.980976178007637       0.20000000000002974      323.58694805436954
    31.999999999999957       0.20000000000003018      323.58694573460093];
T1 = array2table(M, 'VariableNames',{'X','Y','T (K)'})

T1 = 3×3 table
      X          Y       T (K) 
    ______    _______    ______

        32    0.19039    323.57
    31.981        0.2    323.59
        32        0.2    323.59

N = 6;
xv = linspace(min(T1.X), max(T1.X), N);
yv = linspace(min(T1.Y), max(T1.Y), N);
[Xm,Ym] = ndgrid(xv,yv);
TKm = griddata(T1.X, T1.Y, T1.('T (K)'), Xm, Ym);

figure
surfc(Xm, Ym, TKm)
grid on
xlabel('X')
ylabel('Y')
zlabel('T (K)')

There is not much to plot with only 3 data, however if more data are avaialble, this should work with them without modifying the code. Increase ‘N’ to provide effective resolution.

.