CODEWITHSUNDEEP
pythonpandasrolling-computation
wes.e
wes.e

Reputation: 17

How can I generate a conditional rolling average?

I want to be able to perform a rolling average on the Data column based on values which are labeled as True in a separate column. Below is an example data set.

                      Data        Condition
2020-01-01 0:00       2.0          False     
2020-01-01 0:01       2.0          False       
2020-01-01 0:02       732.1        True    
2020-01-01 0:03       734.2        True       
2020-01-01 0:04       733.3        True       
2020-01-01 0:05       731.2        True  
2020-01-01 0:06       2.0          False  
2020-01-01 0:07       2.0          False  
...
2020-06-15 23:52      2.0          False       
2020-06-15 23:53      732.1        True    
2020-06-15 23:54      734.2        True       
2020-06-15 23:55      733.3        True       
2020-06-15 23:56      731.2        True 
2020-06-15 23:57      2.0          False      
2020-06-15 23:58      2.0          False       
2020-06-15 23:59      2.0          False

My expected output, with a window size of 3, for the data column, would be the following:

[2.0, 2.0, 733.15, 733.2, 732.9, 732.25, 2, 2, ..., 2.0, 733.15, 733.2, 732.9, 732.25, 2.0, 2.0, 2.0].

So essentially, for each row, it would take the average of the row above and below that row, as well as the row in question, but always excluding the numbers when it is false.

Here is some code I found, except for that it is not for the current pandas.

pandas.rolling_mean(df[df.Condition], window=30) I found it here: Conditional mean by rolling

Any help or information on this would be greatly appreciated!

Upvotes: 0

Views: 381

Answers (1)

rhug123
rhug123

Reputation: 8768

I believe you would filter your dataframe to only get the rows that fit your condition, then apply a rolling mean.

df['rolling'] = df.loc[df['Condition'] == True]['Data'].rolling(30).mean().fillna(0)

Upvotes: 1

Related Questions