Ad

List Which Doesn't Append In A For Loop In Pandas Autocorrelation

- 1 answer

I have a data Series which looks like this:

            Date         Open         High          Low    Adj Close    Change
4844  26/10/2020  3441.419922  3441.419922  3233.939941  3269.959961 -5.243488
4845  02/11/2020  3296.199951  3529.050049  3279.739990  3509.439941  6.076183
4846  09/11/2020  3583.040039  3645.989990  3511.909912  3585.149902  0.058850
4847  16/11/2020  3600.159912  3628.510010  3543.840088  3557.540039 -1.198015
4848  20/11/2020  3579.310059  3581.229980  3556.850098  3557.540039 -0.611940

I'm trying to create a new list which contains a autocorrelation coefficient for various lookback periods via a for loop. I've tried this:

import pandas as pd
Df = pd.read_csv("SP500 Weekly Data.csv", delimiter=",")
Df.fillna('')
Df['Change'] = ((Df['Adj Close'] - Df['Open']) / Df['Adj Close']*100)
   
for t in range(1,20):
    wk = []
    auto = Df['Change'].autocorr(t).astype(float)
    wk.append(auto)
print(wk)

but instead of getting a list of values, all I get from the print is the last value:

[0.002519726414980291]

At first I thought it was the type of value being returned [I got an ''numpy.float64' object is not iterable' error with .extend()], but .append() doesn't appear to be adding to the list with each loop.

Any help is appreciated, as well as any advice on the mistake I've made, so I can look out for it next time! Thanks

Ad

Answer

In your code, the wk list is being initialized empty every time, Hence you need to place it outside the loop like below for it to work.

import pandas as pd
Df = pd.read_csv("SP500 Weekly Data.csv", delimiter=",")
Df.fillna('')
Df['Change'] = ((Df['Adj Close'] - Df['Open']) / Df['Adj Close']*100)

wk = []   
for t in range(1,20):
    auto = Df['Change'].autocorr(t).astype(float)
    wk.append(auto)
print(wk)
Ad
source: stackoverflow.com
Ad