In epiforecasts/europe-covid-forecast: Forecasting Covid-19 related metrics for the ECDC Forecast Hub

scores <- scoringutils::eval_forecasts(data,
                                         summarise_by = c("model", 
                                                          "location_name", 
                                                          "target_variable",
                                                          "forecast_date"), 
                                       compute_relative_skill = FALSE)


for (loc in locations) {
  cat("\n\n##", loc, "{.tabset} \n\n")

  tmp_scores <- filter(scores, location_name == loc)

  ranked_scores <- tmp_scores %>%
    dplyr::mutate(forecast_date = as.Date(forecast_date)) %>%
    dplyr::group_by(forecast_date, location_name, target_variable) %>%
    dplyr::mutate(num_forecasts = dplyr::n(), 
                  rank = rank(interval_score, ties.method = "average",
                              na.last = "keep"), 
                  standard_rank = round((1 - (rank - 1) / (num_forecasts - 1)) * 100)) %>%
    dplyr::ungroup()

  cat("\n\n### model rank\n\n")
  plot <- score_heatmap(ranked_scores, metric = "rank", 
                        x = "forecast_date",
                        facet_wrap_or_grid = "grid",
                        facet_formula = ~ target_variable) + 
    ggplot2::theme(legend.position = "bottom") +
    ggplot2::scale_fill_gradient(low = "white", high = "coral")

  # suppress duplicated scale fill warning
  suppressWarnings(print(plot))

  cat("\n\n### standardised model rank\n\n")
  plot <- score_heatmap(ranked_scores, metric = "standard_rank", 
                        x = "forecast_date",
                        facet_wrap_or_grid = "grid",
                        facet_formula = ~ target_variable) + 
    ggplot2::theme(legend.position = "bottom") +
    ggplot2::scale_fill_gradient(high = "white", low = "coral")

  # suppress duplicated scale fill warning
  suppressWarnings(print(plot))


}

epiforecasts/europe-covid-forecast documentation built on Jan. 15, 2025, 8:57 p.m.