Yes to be specific, 15 min average of total buy, total sell, and total total.
Suppose we have N sample of total buy data Bi in the last 15 min, the average is
average = sum(Bi) / N
The standard deviation is
Square_root [ sum(Bi2) / N - average2 ]
If these are calculated every time there is a new data point, using data from the last 15 min (counted from this latest data point), we get a moving average and standard deviation. Note that N could change if data come in at irregular intervals.