For column 3, look into diff. diff(x) = [x(2)-x(1), x(3)-x(2), ...
Keep in mind that diff will give you a vector with one less item than in columns 2 and 1. So you may want to say something like:
Column3(1) = NaN;
Column3(2:end) = diff(Column2);
To avoid an error, you need to preallocate the size of Column3. If you haven't already, you need a line before the two above that looks like this:
Column3 = nan(length(Column2), 1);
That creates a variable with the number of items that you need. You can then replace values with the ones that you calculate using diff. (If you preallocate values to be NaN, then you actually don't need the line Column3(1)=NaN because it already is NaN.)
For column 4, look into movmean. The following would give you a moving average with a window width of 7: Column4 = movmean(Column3, 7);
The moving window will be centered on the index of the input (example in comment above). If you want column 4 to represent the average of the last 6 days plus today, you'd want to enter:
Column4 = movmean(Column3, [6, 0]);
Hope this helps!