Hello I have this data set below for 5 years of data. I'm including just a sample.

I would like to take find the average between two dates. For example the average between April 23 and March 26 should be 1.35%. Its important to note it needs to be between two dates not a rolling avg.

I have tried to average between the two periods similar like a y/y change calculation but it gives me an error. The difference between the rows will be constant 52 rows. So I need to find the average between row 52 and row 1.

Y/Y Change Calculation

df=df.pct_change(periods = -52)
df=df.mean(periods=-52) #Error# 

Thank you in advanced for your help.

CodePudding user response：

We can use shift function to get the value from previous periods row and then average the previous value with current value.

df = pd.DataFrame({
    'value':[1,2,3,4,5,6,7,8,9,10]
})
df['lag_3_value'] = df['value'].shift(periods = 3)
df['avg_between_current_and_lag_3'] = (df['value']   df['lag_3_value']) / 2

Output:

CodePudding user response：

Maybe not the shortest line possible, But it is working:

np.mean(df.loc[df['date'].isin(['date_one','date_two'])].value.to_numpy())

Where 'value' is the column name of 'europe'

It will work also given any number of dates in the isin() function.