Area under curve of a dataframe column-CodePudding

I'm trying to calculate the cumulative AUC of a dataframe values from first row to the current row.

Ex:

	points	AUC
0	0	0
1	1	0.5
2	2	1
3	3	4.5
4	4	8
5	5	12.5
6	4	17
7	0	19
8	-2	18
9	-2	16

I can use np.trapz() but I have to calculate it row by row, by a for loop.

for i in df.index:
    row={"AUC" : trapz(df["points"].iloc[:i])}
    df["AUC"].iloc[i]=row

Is there any way to apply it to the whole column without using a for loop?

The second problem is that my dataframe gets updated every minutes so either I have to calculate this cumulative AUC from the beginning of the df which makes the calculation longer and longer, or choose a part of the df (ex: df.tail(25)) and apply a function to it, and by doing this I would lose calculate AUC of the curve before iloc[-25].

CodePudding user response：

I would try something like this:

np.cumsum(df.points)-np.concatenate(([0], np.cumsum(np.diff(df.points)/2)), axis=0)

here is a working example: https://abstra.show/dezL0ASX4s