CODEWITHSUNDEEP
pythonpandas
Stacey
Stacey

Reputation: 5107

Create rolling variance from dataframe column

I have a data-frame (df) which looks like:

              A     B         C          D      
date                                       
2014-02-25  2.41  1.0   20712.0  324.90415  
2014-02-26  2.41  1.0   11400.0  324.90415   
2014-02-27  2.40  1.0    4370.0  323.55600   
2014-02-28  2.37  1.0   51943.0  319.51155   
2014-03-03  2.40  1.0   27902.0  323.55600   
:   :       :       :
2015-03-16  2.39  1.0   18059.0  322.20785   
2015-03-17  2.40  1.0   24346.0  323.55600  
2015-03-18  2.35  1.0  344171.0  316.81525     
2015-03-19  2.35  1.0   15247.0  316.81525   
2015-03-20  2.35  1.0   90217.0  316.81525

I would like to calculate the rolling variance for Column A for the last 5 dates (2015-03-16 - 2015-03-20) looking back at the last 260 days for each calculation.

I can see to calculate variance I can use df['A'].var(). I could create a new data-frame that would just have the last 260 days but was wondering if there was a more elegant way to calculate the 260 days variance for the last 5 dates in the data-frame for column A?

Upvotes: 0

Views: 545

Answers (1)

Quang Hoang
Quang Hoang

Reputation: 150785

Try: 

variance = df.A.rolling(260).var()

which gives you variances for the whole data. Then you can do

variance[-5:]

Upvotes: 1

Related Questions