histogram2() creates a bivariate histogram plot that you can apply to your data.
Here's a demo that applies this plot to noisy gaussian curves. Pay attention the variables x and y which will be inputs to histogram2.
Produce the noisy data. You'll already have the data; you just need to make sure it has the same general shape as these x and y variables.
% Produce noise gaussian data
% x is a 1xm or mx1 vector that defines the x values for each curve.
% y is a mxn matrix of m y-values for n curves.
gausFcn = @(X,C,A,S)exp(-(X - C).^2/(2*S^2)) * A; % guassian function (x, center, amp, sigma)
x = 0 : 0.5 : 3000; % The x values for all curves
nCurves = 20; % Number of curves to produce
y = zeros(numel(x),nCurves); % We'll store the y data here
% Produce n noisy curves
for i = 1:nCurves
y(:,i) = gausFcn(x,randi(100)+1200,(rand(1)+1)*4+10,randi(120)+500) + sin(linspace(0,randi(20)*pi,numel(x)));
end
% Show the noisy data
figure()
sph(1) = subplot(2,1,1);
plot(x,y)
grid on
Apply the bivariate histogram
% First, replicate the x values so there's 1 x value for each y value.
xRep = repmat(x, 1, nCurves); % for row vectors; if col vec: repmat(x, nCurves, 1)
% apply the bivariate historgram
sph(2) = subplot(2,1,2);
linkaxes(sph)
h = histogram2(xRep(:),y(:),'DisplayStyle','tile','ShowEmptyBins','on');
Result
Note, if you'd like it to look like the image you shared, set colormap to gray but reverse the order of colors.
colormap(flipud(gray(255)))
