How to compute the standard deviation of each column in a matrix?

I have some code that creates a matrix of data points and then computes the average value for each column and plots these average values. However, I want to add some error bars to my graph where each error bar represents the standard deviation across each column entry in the original matrix. Is there a concise way to compute the standard deviation for each column so that I can store it and use these standard deviations to make a plot (using something like 'errorbar(x, y, err)'? Thanks!

Answer 1

[the cyclist]

If A is your array, then

s = std(A)

is a row vector, where each element is the standard deviation of the corresponding column. See the documenation for std for details.

Also, be sure to be clear about whether your error bars represent the standard deviation or the standard error of the mean.

[Victoria Helm]

Thanks so much this is super helpful! I decided I also want to try plotting with the SEM for my error bars. I understand that I will have to divide the standard deviation by the square root of the number of elements in each column. However, some of my column elements are NaN so I used std(A, 'omitnan') initially. Is it possible to compute SEM and use 'omitnan'? (I don't know ahead of time how many empty data points I have in each column.

[the cyclist]

You could use the isnan function to keep track of how many elements are NaN. So, calculate the standard deviation, and divide by the square root of the number of elements that are not NaN.

Answer 2

[Steven Lord]

To be certain that you're taking the std over each column, specify the dim input argument as described on the documentation page for the std function.