Creating a heatmap to visualize denisity of 2D point data

Hello,

I am trying to create a heat map from an Mx2 matrix of point data. The point data represents spatial locations and I am attempting to create a heat map that highlights densely-clustered points from sparsely-clustered points. the data is stored in a variable called points. points(:,1) is x data and points(:,2) is y data. when I type HeatMap(points) I get useless information. Is there a way I can visualize the density of these points in a heat map? I tried hist3, but it doesn't represent the data the way I would like.

Thanks,

Kyle

Answer 1

[Walter Roberson]

grid = 256;   %refinement of map

minvals = min(points);
maxvals = max(points);
rangevals = maxvals - minvals;
xidx = 1 + round((points(:,1) - minvals(1)) ./ rangevals(1) * (grid-1));
yidx = 1 + round((points(:,2) - minvals(2)) ./ rangevals(2) * (grid-1));
density = accumarray([yidx, xidx], 1, [grid,grid]);  %note y is rows, x is cols
imagesc(density, 'xdata', [minvals(1), maxvals(1)], 'ydata', [minvals(2), maxvals(2)]);

(This will make the image slightly larger than would be correct. It's probably not worth correcting for.)

[Kyle]

This sort of works. For some reason, the min y is on the top and the max y is on the bottom. This might have something to do with the fact that the ydata has negative numbers in it.

[Kyle]

Got it! All I needed to do was this: set(gca,'YDir','normal') after your code. Thanks!

[Cai Chin]

Hi, I am trying to do a very similar thing - I used your code but it only generates a blue frame instead of a colour density map. Instead of having a matrix input, I have 2 vectors 'v' and 'w' which I tried making into a matrix to fit your code.

I was wondering if you wouldn't mind please pointing out what the issue is or suggesting an alternative?

Thanks in advance.

% Convert vectors 'v' and 'w' into a matrix
points = [v, w];

% Generate colourmap of density
grid = 256;   %refinement of map
minvals = min(points);
maxvals = max(points);
rangevals = maxvals - minvals;
xidx = 1 + round((points(1) - minvals(1)) ./ rangevals(1) * (grid-1));
yidx = 1 + round((points(2) - minvals(2)) ./ rangevals(2) * (grid-1));
density = accumarray([yidx, xidx], 1, [grid,grid]);  %note y is rows, x is cols
imagesc(density, 'xdata', [minvals(1), maxvals(1)], 'ydata', [minvals(2), maxvals(2)]);
set(gca,'YDir','normal');