Rolling Mean/Average within a For Loop on a Dataframe Python

I went through a bunch of posts and couldn't find a more "python" appropiate solution. What I have is a dataframe, then I run a For Loop to calculate several metrics. Throughout the loop many of the columns are dependent on each other so I want to calculate everything up until that point. The issue is that the only way to make the rolling method work within the loop is to run it (or so I think) for the entire column every iteration. I am sure there has to be a better way. Here is a Sample, I have the following DF, where the Column Value is generated within a For Loop:

Minute  Value Rolling Mean
1         3      0
2         5      0
3         8      5.3333
4         4      5.6667
5         6      6
6         7      5.6667

what I am using to calculate the mean through the for loop is this:

n_periods = 3
df['Rolling Mean'] = df['Value'].rolling(n_periods, min_periods = 0).mean(skipna=False)

The proble here is that as you iterate within the for loop it reruns the entire column every row, and I have thousands of them, so it is very slow.
I would want something more like this (which doesn't work), which would only run one calculation per row throughout the loop.

for i in range(1, len(df)):
    df.at[i, 'Rolling Mean'] = df.at[i, 'Value'].rolling(n_periods, min_periods=0).mean(skipna=False)

Any thoughts?
Thank you all!

Ok, so after some trial and error, I figured this line worked. In case anybody else needs it:

# Loop through the indices of the DataFrame
for i in range(1, len(df)):
    # Calculate the rolling mean using df.at and assign it to the 'Rolling Mean' column
    df.at[i, 'Rolling Mean'] = df['Value'][0:i+1].rolling(window=n_periods, min_periods=0).mean().iloc[-1]

Thank you all for the support! this community is amazing!

Maybe it's better, if you try to understand what a rolling mean is and where you want it centered. Then you just implement the math:

for i in range(1, len(df)):
    df.at[i, 'Rolling Mean'] = df.loc[:i, 'Value'].tail(n_periods).mean()

• This needs much less computing power than your answer because it doesn't start from the beginning of your DataFrame.
• The window is right-aligned just like in your example. In general, one can pass center=True to df.rolling but that doesn't work if you don't know the future values. You should be aware of that difference.