macrosynergy.panel.imputers#

impute_panel(df, cids, xcats, threshold=0.5, start_date=None, impute_empty_tickers=False)[source]#

Imputes missing values for each category in a long-format panel dataset by a cross- sectional mean, conditional on the number of available cross-sections at each concerned date exceeding a fraction threshold of the total number of cross- sections.

Parameters:

  • df (DataFrame) – the long-format panel dataset

  • cids (list) – the list of cross sections to be considered in the imputation

  • xcats (list) – the list of categories to be imputed

  • threshold (float) – the fraction of available cross-sections at each date

  • start_date (str) – the starting date for the imputation

  • impute_empty_tickers (bool) – boolean flag for whether to impute missing values for empty tickers

Returns:

the imputed long-format panel data with columns

Return type:

DataFrame

Note

This class is still experimental: the predictions and the API might change without any deprecation cycle.