Skip to contents

Scenario to pick the best scenario out of a list of scenarios

Source: R/scenarios_specific_scenarios.R
scenario_best_of.Rd

scenario_best_of picks the best value between multiple scenarios specified in scenario_names. small_is_best allows to pick if higher value or smaller values are best.

Usage

scenario_best_of(
  df,
  scenario_names,
  value_col = "value",
  maximize_end_year = TRUE,
  start_year = 2018,
  end_year = 2025,
  target_year = end_year,
  scenario_name = glue::glue("best_of_{paste0(scenario_names, collapse = '_')}"),
  scenario_col = "scenario",
  trim = TRUE,
  small_is_best = FALSE,
  keep_better_values = TRUE,
  upper_limit = 100,
  lower_limit = 0,
  trim_years = TRUE,
  start_year_trim = start_year,
  end_year_trim = end_year,
  ind_ids = billion_ind_codes("all")
)

Arguments

df

Data frame in long format, where 1 row corresponds to a specific country, year, and indicator.

scenario_names

names of the scenario to pick from.

value_col

Column name of column with indicator values.

maximize_end_year

(Boolean) if TRUE, the best scenario is picked on the best value at end_year. Default to FALSE.

start_year

Start year for scenario, defaults to 2018.

end_year

End year for scenario, defaults to 2025

target_year

Year by which the scenario should eventually be achieved. Defaults to end_year

scenario_name

Name of the scenario. Defaults to scenario_percent_change_baseline_year

scenario_col

Column name of column with scenario identifiers. Useful for calculating contributions on data in long format rather than wide format.

trim

logical to indicate if the data should be trimmed between upper_limit and lower_limit.

small_is_best

Logical to identify if a lower value is better than a higher one (e.g. lower obesity in a positive public health outcome, so obesity rate should have small_is_best = TRUE).

keep_better_values

logical to indicate if "better" values should be kept from value_col if they are present. Follows the direction set in small_is_best. For instance, if small_is_best is TRUE, then value_col lower than col will be kept.

upper_limit

limit at which the indicator should be caped. Can take any of "guess", or any numeric. guess (default) will take 100 as the limit if percent_change is positive, and 0 if negative.

lower_limit

limit at which the indicator should be caped. Can take any of "guess", or 0 to 100. guess (default) will take 0 as the limit if percent_change is positive, and 100 if negative.

trim_years

logical to indicate if years before start_year_trim and after end_year_trim should be removed

start_year_trim

(integer) year to start trimming from.

end_year_trim

(integer) year to end trimming.

ind_ids

Named vector of indicator codes for input indicators to the Billion. Although separate indicator codes can be used than the standard, they must be supplied as a named vector where the names correspond to the output of billion_ind_codes().

Details

If multiple scenario are tied for the best scenario, the scenario with the best time series for each iso3 and ind is picked. If they are still tied scenarios for the best value, the first in alphabetical order is picked.

See also

Basic scenarios scenario_aroc(), scenario_bau(), scenario_linear_change(), scenario_percent_baseline()

Comparing scenarios scenario_bau(), scenario_best_in_region(), scenario_quantile()