OR another option is:

df %>%
  tidyr::pivot_longer(cols = c(sumCumDeaths, sumCumRecovered, sumCumTests),
                      names_to = "variable",
                      values_to = "value") %>%
  dplyr::group_by(Continent, variable) %>%
  dplyr::mutate(variable = fct_relevel(variable, c("Africa", "Asia", "Europe", "North America", "Oceania", 
                                            "South America"))) %>%
  ggplot(ggplot2::aes(variable, y = log(value), fill=Continent))  
  geom_col(position = position_dodge())