How can I add a summation column to a heatmap?

Hi everyone,

I am trying to make a plot with bins of wave heights on one axis and wave periods on the other, with each bin colored by its percent occurrence over a period of time in my dataset. I was able to "trick" MATLAB's heatmap function into doing this with a dummy variable for the Color Data, but I have not been able to figure out how to add a row and column summing up the percentages of each respective row and column. Below is the figure that I have made:

What I would like to do is something like the following (which I think is just a table in Excel):

Note the SUM [%] row and column; I would like to add these to the former plot. I have looked through the documentation for the heatmap function and can't seem to find anything. Let me know if you have any ideas! Thanks.

Answer 1

[Voss]

Augment your matrix with the row and column sums before sending it to heatmap. Then set the color-limits to the range of your original matrix.

data = magic(5);
data_limits = [min(data(:)) max(data(:))];
data(end+1,:) = sum(data,1);
data(:,end+1) = sum(data,2);
heatmap(data);
caxis(data_limits) % set the color-limits

You may get some variation in the color in the sum column and sum row, in general.

[Chase]

Thanks for the response. This is a helpful trick! It would work great, except that I am using heatmap to calculate my Color Data. So I guess now my problem is figuring out how to do what heatmap does with my own code! Maybe you can help with that?

What I want to do is count the number of times each unique column of bins occurs:

wave_heights = [1 4 10 5 3 5 6 10 7 6];
wave_periods = [2 5 6 3 5 6 5 4 5 4];

edges_wvh = fix(min(wave_heights)):1:ceil(max(wave_heights));
edges_wvp = fix(min(wave_periods)):0.5:ceil(max(wave_periods));

wvh_discretize = discretize(wave_heights,edges_wvh,'categorical')

wvh_discretize = 1×10 categorical array
     [1, 2)      [4, 5)      [9, 10]      [5, 6)      [3, 4)      [5, 6)      [6, 7)      [9, 10]      [7, 8)      [6, 7) 

wvp_discretize = discretize(wave_periods,edges_wvp,'categorical')

wvp_discretize = 1×10 categorical array
     [2, 2.5)      [5, 5.5)      [5.5, 6]      [3, 3.5)      [5, 5.5)      [5.5, 6]      [5, 5.5)      [4, 4.5)      [5, 5.5)      [4, 4.5) 

Using the above example, I'd hope to get out something that counts the number of times the combination of wave height bin [1, 2) and wave period bin [2, 2.5) occurs, then number of times the combination of wave height bin [4, 5) and wave period bin [5, 5.5) occurs, and so on till all unique columns have been counted. Then I could plot the respective percentage counts for each combination as my color data and include a sum column in my color data matrix like you demonstrated above.

[Voss]

Sounds like histcounts2 may be useful in this situation.

wave_heights = [1 4 10 5 3 5 6 10 7 6];
wave_periods = [2 5 6 3 5 6 5 4 5 4];

edges_wvh = fix(min(wave_heights)):1:ceil(max(wave_heights));
edges_wvp = fix(min(wave_periods)):0.5:ceil(max(wave_periods));

histcounts2(wave_heights,wave_periods,edges_wvh,edges_wvp)

ans = 9×8

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1

[Chase]

Thanks, I didn't know about histcounts2! I ended up using groupcounts, which did basically the same thing.

[Voss]

You're welcome!