Measure of how well two dataframe columns move with each other-CodePudding

I have two columns in a df and rows of dates. I'd like to see how well each column matches the other and moves in sync with the other column - ie. do they move in tandem and does one influence the movements in the other.

                                  Col1      Col2     
Date                                     
1991-01-01 00:00:00 00:00        6.945847   3.4222
1991-04-01 00:00:00 00:00        8.377481   6.7783
1991-07-01 00:00:00 00:00        7.869787   4.6666
...                                   ...

Is there a way to do this in pandas?

I thought of dividing each row by the value in the first row to see the % increase, but wondered if there was a better statistical way of doing this.

Thanks.

CodePudding user response：

If you want to calculate Spearman correlation coefficient you can use Scipy package

df = pd.DataFrame({'Date': ['1991-01-01 00:00:00 00:00', '1991-04-01 00:00:00 00:00', '1991-07-01 00:00:00 00:00'], 
                  'Col1': [6.945847 , 8.377481, 7.869787], 
                  'Col2': [3.4222, 6.7783, 4.6666]}).set_index('Date')

from scipy import stats
stats.spearmanr(df['Col1'], df['Col2'])
>>>
SpearmanrResult(correlation=1.0, pvalue=0.0)

CodePudding user response：

https://en.wikipedia.org/wiki/Cross-correlation

https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.correlate.html

Spearman only tells you the correlation at zero shift.