# Install coursekata if not already installed
if (!require("coursekata", quietly = TRUE)) {
    install.packages("coursekata", repos = "https://cloud.r-project.org")
}

suppressPackageStartupMessages({
    library(coursekata)
    library(dplyr)
    library(lubridate)
    library(tidyr)
})

# Load data from the same folder as this notebook
df <- read.csv("fifa_players.csv")

# Check dimensions
dim(df)

# Look at variable names
names(df)

# Look at first few rows
head(df[, c("name", "birth_date", "nationality", "positions")], 10)

# Convert birth_date to date format and extract month
df <- df %>%
  mutate(
    birth_date_parsed = as.Date(birth_date, format = "%m/%d/%Y"),
    birth_month = month(birth_date_parsed)
  )

# Check a few examples
head(df[, c("name", "birth_date", "birth_month")], 10)

# Verify we have all 12 months
table(df$birth_month)

# Count players by birth month
tally(~ birth_month, data = df)

# Create a summary table with month names
month_names <- c("January", "February", "March", "April", "May", "June",
                 "July", "August", "September", "October", "November", "December")

birth_month_counts <- df %>%
  group_by(birth_month) %>%
  summarize(count = n()) %>%
  mutate(month_name = month_names[birth_month]) %>%
  arrange(birth_month)

birth_month_counts

# Create month name variable for better labels
df <- df %>%
  mutate(month_name = factor(month_names[birth_month], levels = month_names))

# Bar chart showing counts
gf_bar(~ month_name, data = df) %>%
  gf_labs(x = "Birth Month", y = "Number of Players",
          title = "Distribution of Birth Months Among Professional Soccer Players")

# Bar chart showing proportions
gf_props(~ month_name, data = df) %>%
  gf_labs(x = "Birth Month", y = "Proportion of Players",
          title = "Proportion of Players by Birth Month")

# Calculate expected count per month (uniform distribution)
total_players <- nrow(df)
expected_per_month <- total_players / 12

cat("Total players:", total_players, "\n")
cat("Expected players per month (if uniform):", round(expected_per_month, 2), "\n")

# Add expected values to our summary
birth_month_summary <- df %>%
  group_by(birth_month, month_name) %>%
  summarize(observed = n(), .groups = 'drop') %>%
  mutate(expected = expected_per_month,
         difference = observed - expected)

birth_month_summary

# Prepare data for comparison plot
comparison_data <- birth_month_summary %>%
  select(month_name, observed, expected) %>%
  pivot_longer(cols = c(observed, expected), 
               names_to = "type", 
               values_to = "count")

# Create comparison plot
gf_col(count ~ month_name, fill = ~ type, data = comparison_data, position = "dodge") %>%
  gf_labs(x = "Birth Month", y = "Number of Players", fill = "Type",
          title = "Observed vs. Expected Birth Month Distribution") %>%
  gf_hline(yintercept = ~ expected_per_month, linetype = "dashed", color = "gray")

# Show differences sorted
birth_month_summary %>%
  arrange(desc(difference)) %>%
  select(month_name, observed, expected, difference)

# Create Italy vs. Other comparison
df <- df %>%
  mutate(country_group = ifelse(nationality == "Italy", "Italy", "Other Countries"))

# Count by country group and month (include birth_month to use later)
italy_comparison <- df %>%
  group_by(country_group, month_name, birth_month) %>%
  summarize(count = n(), .groups = 'drop') %>%
  group_by(country_group) %>%
  mutate(proportion = count / sum(count)) %>%
  ungroup()

# Show summary
italy_comparison %>%
  group_by(country_group) %>%
  summarize(
    jan_mar = sum(proportion[birth_month %in% 1:3]),
    oct_dec = sum(proportion[birth_month %in% 10:12]),
    .groups = 'drop'
  )

# Visualize with faceted plot
gf_props(~ month_name, data = df) %>%
  gf_facet_grid(. ~ country_group) %>%
  gf_labs(x = "Birth Month", y = "Proportion of Players",
          title = "Birth Month Distribution: Italy vs. Other Countries")

# Extract primary position (first position listed)
df <- df %>%
  mutate(
    primary_position = case_when(
      grepl("GK", positions) ~ "Goalkeeper",
      grepl("CB|LB|RB|LWB|RWB", positions) ~ "Defender",
      grepl("CM|CAM|CDM|LM|RM", positions) ~ "Midfielder",
      grepl("ST|CF|LW|RW", positions) ~ "Forward",
      TRUE ~ "Other"
    )
  )

# Filter to main positions and visualize
df_main_positions <- df %>%
  filter(primary_position != "Other")

gf_props(~ month_name, data = df_main_positions) %>%
  gf_facet_grid(. ~ primary_position) %>%
  gf_labs(x = "Birth Month", y = "Proportion of Players",
          title = "Birth Month Distribution by Position")

# Calculate Q1 vs Q4
quarter_summary <- df %>%
  mutate(
    quarter = case_when(
      birth_month %in% 1:3 ~ "Q1 (Jan-Mar)",
      birth_month %in% 4:6 ~ "Q2 (Apr-Jun)",
      birth_month %in% 7:9 ~ "Q3 (Jul-Sep)",
      birth_month %in% 10:12 ~ "Q4 (Oct-Dec)"
    )
  ) %>%
  group_by(quarter) %>%
  summarize(
    count = n(),
    proportion = n() / nrow(df),
    .groups = 'drop'
  )

quarter_summary

# Calculate Q1/Q4 ratio
q1_count <- sum(df$birth_month %in% 1:3)
q4_count <- sum(df$birth_month %in% 10:12)
q1_q4_ratio <- q1_count / q4_count

cat("\nQ1 (Jan-Mar) players:", q1_count, "\n")
cat("Q4 (Oct-Dec) players:", q4_count, "\n")
cat("Q1/Q4 ratio:", round(q1_q4_ratio, 2), "\n")
cat("This means there are", round(q1_q4_ratio, 2), "times more Q1 players than Q4 players\n")

# Extract birth year and create decade groups
df <- df %>%
  mutate(
    birth_year = year(birth_date_parsed),
    birth_decade = (birth_year %/% 10) * 10
  )

# Calculate Q1 proportion by decade
decade_effect <- df %>%
  mutate(q1_born = birth_month %in% 1:3) %>%
  group_by(birth_decade) %>%
  summarize(
    total = n(),
    q1_count = sum(q1_born),
    q1_proportion = mean(q1_born),
    .groups = 'drop'
  ) %>%
  arrange(birth_decade)

decade_effect

# Visualize trend
gf_col(q1_proportion ~ factor(birth_decade), data = decade_effect) %>%
  gf_hline(yintercept = 0.25, linetype = "dashed", color = "red") %>%
  gf_labs(x = "Birth Decade", y = "Proportion Born in Q1 (Jan-Mar)",
          title = "Relative Age Effect Over Time",
          subtitle = "Red line shows expected 25% if uniform")