Resampling Multivariate Time Series: Resampling Plugin
Hi, community!
I Have a dataset with 3 columns( Date, Item_ID, quantity), my period of studies is from 2015 to 2020.
Some Item Id don't have a full history, I want to resample the dataset, and I used the TS plugging resampling with these settings:
But I don't get the desired result, did I miss something? Or id there another way to achieve this?
Example of the desired output for this Item Id:
Best Answer
-
Hey Dataikers!
I managed to solve the problem by fixing the day in the input dates to the last day of the month instead of the first day of the month I used before, and then the sampling works just fine.
Thank you for your feedback
Answers
-
Welcome to the Dataiku Community.
For a bunch of additional information about time series preparation in the following Dataiku academy lesson.
https://academy.dataiku.com/path/ml-practitioner/time-series-preparation-1/609907
@CoreyS
the Academy seems to have gone down just now. -
Must have been a momentary blip because its working for me!
-
OK, it's back up and running for me as well. Thanks.
-
Hey!
Thanks for your feedback, I checked the resources, the problem for me was that the format of the date column was:
2011-01-01T00:00:00.000Z
2013-01-01T00:00:00.000Z
2015-01-01T00:00:00.000Z
And when applying the sampling, with the 'Years' unit, I expected that only the year will be incremented, while the output also fixes the month to 12 and the day to 31.
I want to know in which format should I put my date column before the sampling, if I would like to use the 'Months' unit?
