Skip to main content
H
Seasoned ⭐️⭐️⭐️
henning.kvinnesland
Seasoned ⭐️⭐️⭐️
January 19, 2022
Solved

Fallback for missing values when calculating synthetic timeseries

  • January 19, 2022
  • 2 replies
  • 231 views

Hi!

 

If one or more of the source-timeseries are missing data-points they will also be missing in the synthetic timeseries output..

We would like an option to default to zero for missing values so that the API will use the available values to calculate a partial result.

 

Example:

TS1: [1,2,3]

TS2: [3, undefined, 4]

 

Output now if calculating a sum of TS1 + TS2:

[4, undefined, 7]

 

What we would like:

[4, 2, 7] 

 

    Best answer by Ola Øyan

    Hi! 

    Ok, a bit of a misunderstanding on our side here, so to make sure you understand it correctly, I am providing some more information. The issue is not meant to address gaps in the data, but rather timeseries of different 'lengths'. I see that these probably gets muddled together in practice, but the case for the feature described is different: If fx Trafo B was put into service in 2021, affecting the load on Trafo A that has been running since 2011. The analyst want to make a calculation of what the situation was like for Trafo A +B before and after. The current way to add these together in our app would only give data from 2021, where both timeseries have data. What we want with the checkbox is to 'extend' the timeseries for Trafo B, so that we get a result where the grid change can be understood (see our design suggestion below). How this relates to gaps in the data is another question, but I hope this is more clear..

     

    2 replies

    T
    Expert ⭐️⭐️⭐️⭐️
    Thomas Sjolshagen
    Lead Product Manager
    January 21, 2022

    Hi and thank you for the insight @henning.kvinnesland !

     

    I’ve added this problem as “Product Feedback”  here on the Hub.

    Please upvote and provide any additional commentary you may have there, so we can be sure to include the information in our discovery and assessment process for Time Series and Synthetic Time Series.

      O
      Committed ⭐️⭐️⭐️
      Ola ØyanAnswer
      Committed ⭐️⭐️⭐️
      January 25, 2022

      Hi! 

      Ok, a bit of a misunderstanding on our side here, so to make sure you understand it correctly, I am providing some more information. The issue is not meant to address gaps in the data, but rather timeseries of different 'lengths'. I see that these probably gets muddled together in practice, but the case for the feature described is different: If fx Trafo B was put into service in 2021, affecting the load on Trafo A that has been running since 2011. The analyst want to make a calculation of what the situation was like for Trafo A +B before and after. The current way to add these together in our app would only give data from 2021, where both timeseries have data. What we want with the checkbox is to 'extend' the timeseries for Trafo B, so that we get a result where the grid change can be understood (see our design suggestion below). How this relates to gaps in the data is another question, but I hope this is more clear..

       

        Terms & ConditionsAccessibility statement