The Data

The data this week comes from Benjamin Nowak by way of International Trail Running Association (ITRA). Their original repo is available on GitHub.

race <- read_csv("race.csv") %>%
  mutate(country_code = 
           countrycode(country, 'country.name', 'iso3c')) %>%
  mutate(participation = ifelse(participation == "Solo", "solo", participation)) %>%
  filter(participation == "solo")


ultra_rankings <- read_csv("ultra_rankings.csv")

Home Field Advantage?

ultra_rankings <- ultra_rankings %>%
  left_join(race, by = "race_year_id")
country_ranks <- ultra_rankings %>%
  mutate(advantage = ifelse(country_code == nationality, "home", "away")) %>%
  group_by(country_code, advantage, gender) %>%
  summarize(ave_rank = mean(rank, na.rm = TRUE),
            top_rank = min(rank, na.rm = TRUE),
            n_group = n()) %>%
  filter(!is.na(country_code)) %>%
  filter(!is.na(gender)) %>%
  pivot_wider(id_cols = c(country_code, gender),
              names_from = advantage,
              values_from = c(ave_rank, top_rank, n_group))
library(ggalt)
country_ranks %>%
  filter(!is.na(n_group_away) & !is.na(n_group_home)) %>%
  filter(n_group_away > 50 & n_group_home > 50) %>%
  mutate(home_adv = ifelse(ave_rank_home < ave_rank_away, "yes", "no")) %>%
  ggplot() + 
  geom_dumbbell(aes(y = country_code, 
                    x=ave_rank_away,
                    xend = ave_rank_home,
                    colour = home_adv),
                    colour_xend = "green",
                    colour_x = "purple") + 
  facet_wrap(~gender)

country_ranks %>%
  filter(!is.na(n_group_away) & !is.na(n_group_home)) %>%
  filter(n_group_away > 50 & n_group_home > 50) %>%
  mutate(home_adv = ifelse(top_rank_home < top_rank_away, "yes", "no")) %>%
  ggplot() + 
  geom_dumbbell(aes(y = country_code, 
                    x=top_rank_away,
                    xend = top_rank_home,
                    colour = home_adv),
                    colour_xend = "green",
                    colour_x = "purple") + 
  facet_wrap(~gender)

Home

What proportion of winners are from the race’s country?

country_wins <- ultra_rankings %>%
#  mutate(advantage = ifelse(country_code == nationality, "home", "away")) %>%
  filter(rank == 1) %>%
  group_by(country_code) %>%
  mutate(n_runs = n()) %>%
  ungroup() %>%  
  group_by(country_code, gender) %>%
  summarize(prop_win = mean(nationality == country_code, na.rm = TRUE),
            n_group = n(),
            n_runs = sample(n_runs, 1)) %>%
  filter(!is.na(country_code)) %>%
  filter(!is.na(gender)) #%>%
  #pivot_wider(id_cols = c(country_code, gender),
  #           names_from = advantage,
  #            values_from = c(prop_win, n_group))
library(forcats)

country_wins %>%
  mutate(prop_loss = 1 - prop_win) %>%
  mutate(home_adv = ifelse(prop_win < 0.5, "no", "yes")) %>%
  filter(n_runs > 10) %>%
  ggplot() + 
  geom_col(aes(y = country_code, 
               x = prop_win, color = home_adv),
           fill = "lightgrey") + 
  facet_wrap(~gender) +
  scale_color_manual("Home Advantage?", values = c("#5c1a33", "Dodgerblue"), label = c("no", "yes")) +
  labs( x = "",
        y = "",
        title = "Proportion of Wins by Resident Runner",
        caption = "Tidy Tuesday Plot: @hardin47 | Data: Ultra Trail Running",
        legend = "Home Advantage?")
Broken down by country, a bar plot measuring how often the race is won by a runner whose nationality is the same as the location of the race. The majority of the races are won by a resident well over 50% of the time.

How often is the ultra trail run won by a runner whose nationality is the same as the location of the race? 

praise()
## [1] "You are phenomenal!"