search

HichamJohra / energy_flexibility_kpis

Energy flexibility KPIs
MIT License
5 stars 6 forks source link

Provide function for resampling profiles #15

Open kingsleynweye opened 1 year ago

kingsleynweye commented 1 year ago

Some KPIs are time resolution agnostic and require the user to provide profiles in the resolution that they want KPIs to be calculated in. To help with profile preparation, the energy_flexibility_kpis.utilities will have a resampling function.

HichamJohra commented 1 year ago

Perfect :) All those "support" function for data pre-treatment before direct use into KPIs function will thus be placed in "utilities" I guess

kingsleynweye commented 1 year ago

Perfect :) All those "support" function for data pre-treatment before direct use into KPIs function will thus be placed in "utilities" I guess

Yes, that's the plan :).

HichamJohra commented 1 year ago

@kingsleynweye , @ZhelunChen : another nice set of "utilities" functions could some to check the quality of the input dataset. Something like a set of basic functions:

this functions would give information to the user when uploading data before calculating KPIs. We could set some warning message for pre-defined data quality thresholds.

What do you think?

kingsleynweye commented 1 year ago

@kingsleynweye , @ZhelunChen : another nice set of "utilities" functions could some to check the quality of the input dataset. Something like a set of basic functions:

  • give basic intel on the time series: mean, median, 10th/90th percentile, total length in time, sampling distribution/time step distribution, total number of data points, number of NaN
  • quality data function: take "sampling rate" as input and check the number of missing data points according to that sampling rate, gives distribution of data gaps (occurrence data gaps as a function of length of data gaps), overall % of missing information.

this functions would give information to the user when uploading data before calculating KPIs. We could set some warning message for pre-defined data quality thresholds.

What do you think?

@HichamJohra yes, I agree with providing a utility functions that return summary statistics. I like the quality check and using that to provide warnings.

I'll create a new issue concerning this.