compute_width: Summarize forecast interval widths
In dshemetov/evalcast-mirror: Tools For Evaluating COVID Forecasters

Summarize forecast interval widths

compute_width(
  predictions_cards,
  grp_vars = c("forecaster", "forecast_date", "ahead"),
  avg_vars = c("geo_value"),
  levels = c(0.5, 0.7, 0.9)
)

`predictions_cards`	predictions cards (or a data frame with columns `quantile`, `value` and any grouping vars and averaging vars)
`grp_vars`	variables over which to compare widths
`avg_vars`	variables over which we compute quantiles.
`levels`	Quantile levels to summarize the interval widths. For example, `levels = c(0.5, 0.7, 0.9)`, the default, computes the median, 70% and 90% quantiles of interval widths within each combination of `grp_vars` over the `avg_vars`