library(tidyverse)


# --- load athlete data

# We are only interested in following disciplines
std_runs <- c("100 metres", "200 metres", "400 metres", "800 metres", 
              "1,500 metres", "5,000 metres", "10,000 metres", "Marathon")

read_csv("athlete_events.csv") %>% 
  filter(Sport == "Athletics") %>% 
  select(Name, Sex, Age, Height, Weight, Year, Event, Medal) %>% 
  mutate(
    Event = str_remove(Event, "Athletics Women's "),
    Event = str_remove(Event, "Athletics Men's ")) %>% 
  filter(Event %in% std_runs) ->
  athletes


# --- Top Medalists

athletes %>% 
    pivot_wider(names_from = Medal, values_from = Medal) %>%
    mutate(across(c(Gold, Bronze, Silver), ~ !is.na(.))) %>%
    group_by(Name) %>%
    summarize(across(where(is.logical), sum)) %>%
    arrange(-Gold, -Silver, -Bronze) %>%
    transmute(Name, Gold, Silver, Bronze) ->
    medal_ranking


# --- Age distribution among Women's Marathon participants

medal_color <- c(Bronze = "#6A3805", Silver = "#B4B4B4", Gold = "#AF9500")

athletes %>% 
    filter(Event == "Marathon", Sex == "F", Year > 1980) %>%
    mutate(Year = as.factor(Year)) ->
    d

ggplot(d, aes(x = Year, y = Age)) +
    geom_boxplot(na.rm=T) +
    geom_point(data = drop_na(d), mapping = aes(color = Medal)) +
    scale_color_manual(values = medal_color) +
    ggtitle("Age distribution among Women's Marathon participants")

# --- Change in height of male runners

athletes %>% 
    mutate(Event = factor(Event, levels=std_runs)) %>% # use factor for ordering
    filter(Sex == "M", Year > 1900) %>%
    group_by(Event, Year) %>%
    summarize(MeanHeight = mean(Height, na.rm=T)) %>%
    ggplot(aes(x = Year, y = MeanHeight, color = Event)) +
        geom_point() + 
        ggtitle("Men's runs - mean across participants") +
        geom_smooth(se = FALSE)

# --- Medalist times

# Times are given as strings with inconsistent format.
# Need custom function for conversion in seconds
str2sec <- function(s) { 
  s %>% 
    str_split(":|h|-") %>% 
    sapply(function(x) {
      v <- as.double(x)
      v3 <- c(0,0,0)
      v3[(4-length(v)):3] <- v
      v3[1] * 3600 + v3[2] * 60 + v3[3]
    })
}

# We are only interested in following disciplines
std_runs <- c("100M", "200M", "400M", "800M", "1500M", "5000M", "10000M", "Marathon")

read_csv("results.csv") %>% 
  mutate(
    Event = str_remove(Event, c(" Men")),
    Event = str_remove(Event, c(" Women"))) %>% 
  filter(Event %in% std_runs) %>% 
  mutate(Result = str2sec(Result)) %>% 
  drop_na() ->
  runs

medal_color <- c(B = "#6A3805", S = "#B4B4B4", G = "#AF9500")

ggplot(runs, aes(x = Year, y = Result, shape = Gender)) +
    facet_wrap(vars(factor(Event, levels = std_runs)), scales = "free_y") +
    geom_point(aes(color = Medal)) +
    scale_color_manual(values = medal_color) +
    ggtitle("Times of medal winners in different running disciplines") +
    xlab("Year") +
    ylab("Time in seconds") +
    geom_smooth(se = T)