pandas rolling correlation with a subset

Question

If I want a rolling correlation of a series with a particular subset, is this possible with pandas rolling method without using for loop? An example would be something like this.

# given this sample dataframe
df = pd.DataFrame(columns = ['a'])
df['a'] = [0.01, 0.02, -0.21, 0.05]

# rolling corr example
df.rolling(10).corr(df.iloc[-10:,:])

Answer 1

• I was able to implement a rolling correlation with this function

def func(x):
    a = [0.002, -0.025, 0.0037]
    return np.corrcoef(x,a)[0,1]

df.rolling(3).apply(func)