20241210 NHL Teams game state.R

# -- Created by Owen Thompson (@OTStats)
# 2024-12-10

# -- Load libraries
library(tidyverse)
library(ggchicklet)
library(ggtext)


# -- Load data
# Data downloaded on 12/11/24 from Moneypuck.com
df <- read_csv("Downloads/shots_2024 3.csv")


# -- Prep data
df_summary <- 
  df %>% 
  filter(isPlayoffGame == 0) %>% 
  group_by(game_id = str_c(season, "-", game_id)) %>% 
  transmute(
    season, 
    game_id, 
    period, 
    home_team = homeTeamCode, 
    away_team = awayTeamCode, 
    home_goals = homeTeamGoals, 
    away_goals = awayTeamGoals, 
    time, 
    game_score = str_c(homeTeamGoals, "-", awayTeamGoals), 
    score_change = if_else(game_score != lag(game_score, default = "1"), 1, 0)
  ) %>% 
  ungroup() %>% 
  filter(score_change == 1) %>% 
  mutate(
    lead_diff = home_goals - away_goals,  # Calculate the lead difference
    home_lead = lead_diff > 0,
    away_lead = lead_diff < 0
  ) %>%
  group_by(game_id) %>%
  mutate(
    winner = if_else(last(home_goals) > last(away_goals), "home", "away"),
    home_behind = any(lag(away_lead, default = FALSE) & winner == "home"),
    away_behind = any(lag(home_lead, default = FALSE) & winner == "away"),
    comeback_win = case_when(
      winner == "home" & home_behind ~ TRUE,
      winner == "away" & away_behind ~ TRUE,
      TRUE ~ FALSE
    )
    # max_home_lead = max(lead_diff[lead_diff > 0], na.rm = TRUE),   # Greatest lead for home team
    # min_home_deficit = min(lead_diff[lead_diff > 0], na.rm = TRUE),   # Greatest lead for home team
    # max_away_deficit = min(lead_diff[lead_diff > 0], na.rm = TRUE)*-1,   # Greatest lead for home team
    # max_away_lead = max(lead_diff[lead_diff > 0], na.rm = TRUE)*-1,   # Greatest lead for home team
  ) %>%
  ungroup()

# -- Table for end of the game
end_game_df <- df_summary %>% 
  filter(time <= 3600) %>% 
  group_by(game_id) %>% 
  filter(time == max(time)) %>% 
  mutate(time = 3600)

# start_game_df <- df_summary %>% 
#   filter(time <= 3600) %>% 
#   group_by(game_id) %>% 
#   filter(time == min(time)) %>% 
#   mutate(time = 0)


df_game_state_summary <- 
  df_summary %>% 
  filter(time <= 3600) %>% 
  bind_rows(end_game_df) %>% 
  arrange(game_id, time) %>% 
  group_by(game_id) %>% 
  mutate(game_score_id = row_number(), .after = "game_id") %>%  
  mutate(time = if_else(game_score_id == 1, 0, time), # prep time for initial 0-0
         time_in_state = lead(time) - time, .after = time) %>% 
  ungroup()
# filter(team != c("L.A", "N.J", "S.J"))



df_prepped_for_viz <- 
  bind_rows(
    transmute(df_game_state_summary, 
              team = home_team, 
              opponent = away_team, 
              game_id, 
              time, 
              time_in_state, 
              goals = home_goals, 
              opp_goals = away_goals), 
    transmute(df_game_state_summary, 
              team = away_team, 
              opponent = home_team, 
              game_id, 
              time, 
              time_in_state, 
              goals = away_goals, 
              opp_goals = home_goals)
  ) %>% 
  mutate(goal_diff = goals - opp_goals, 
         goal_diff_fct = case_when(goal_diff >= 3 ~ "Up 3+",
                                   goal_diff <= -3 ~ "Dn 3+", 
                                   goal_diff == 2 ~ "Up 2", 
                                   goal_diff == 1 ~ "Up 1", 
                                   goal_diff == 0 ~ "Tied", 
                                   goal_diff == -1 ~ "Dn 1", 
                                   goal_diff == -2 ~ "Dn 2", 
                                   TRUE ~ str_c(goal_diff)
         )) %>% 
  filter(!is.na(time_in_state)) %>% 
  mutate(game_state_winning = if_else((goals-opp_goals) > 0, time_in_state, 0), 
         team_order = fct_reorder(team, game_state_winning, .fun = sum)) %>% #filter(is.na(game_state_winning))
  group_by(
    # team, 
    team_order, 
    goal_diff_fct = factor(goal_diff_fct, levels = c("Dn 3+", "Dn 2", "Dn 1", "Tied", "Up 1", "Up 2", "Up 3+"))) %>% 
  summarize(total_time = sum(time_in_state, na.rm = TRUE), 
            total_time_winning = sum(game_state_winning), 
            .groups = 'drop') %>% 
  with_groups(team_order, mutate, overall_time = sum(total_time)) %>% 
  mutate(percent_of_time = total_time/overall_time, 
         game_state_winning = if_else(goal_diff_fct %in% c("Up 1", "Up 2", "Up 3+"), percent_of_time, 0), 
         team_order = fct_reorder(team_order, game_state_winning, .fun = sum))




df_prepped_for_viz %>% 
  group_by(team_order = str_c(team_order)) %>% 
  summarize(time_winning = sum(total_time_winning)) %>% 
  arrange(desc(time_winning))




# source("Downloads/add_logo.R")
# remotes::install_git("https://git.rud.is/hrbrmstr/ggchicklet.git")
# remotes::install_gitlab("hrbrmstr/ggchicklet")
# library(ggchicklet)
# library(ggtext)
df_prepped_for_viz %>% 
  ggplot(aes(x = team_order, 
             y = percent_of_time, 
             fill = goal_diff_fct %>% fct_rev())) + 
  geom_chicklet(position = ggplot2::position_stack(reverse = F)) +
  
  geom_hline(yintercept = .25, linetype = "dashed", colour = "white") + 
  geom_hline(yintercept = .5, linetype = "dashed", colour = "white") + 
  geom_hline(yintercept = .75, linetype = "dashed", colour = "white") + 
  labs(fill = "", 
       x = "", 
       y = "", 
       caption = "Data as of 12/11/2024\nSource: Moneypuck.com\nMotivation: @owenlhjphillips", 
       # title = "How often are teams Up, Down, or Tied", 
       title = "<span style='font-size:20pt'>NHL 2024: How often are teams <span style='color:#738877;'>Up</span>, 
       <br><span style='color:#6A51A3;'>Down</span>, or <span style='color:#B3B3B3;'>Tied</span></span>", 
       subtitle = "Teams are sorted by how often they are winning") + 
  scale_y_continuous(
    expand = c(0, 0.005), 
    breaks = c(0, .25, .5, .75, 1), 
    labels = c("0%", "25%", "50%", "75%", "100%")
  ) + 
  coord_flip() + 
  scale_fill_manual(values = 
                      c(
                        "#475841", 
                        "#738877", 
                        "#9fb8ad", 
                        "grey90", 
                        "#DADAEB", 
                        "#9E9AC8", 
                        "#6A51A3"
                      )) +
  theme(legend.position = 'top',
        legend.direction = "horizontal",
        legend.box = "horizontal",
        legend.text = element_text(face = 'bold'), 
        # legend.text = element_text(hjust = 0.5) 
        plot.title = element_markdown(hjust =0.5, face = "bold"), 
        plot.subtitle = element_text(hjust = 0.5, size = 10), 
        axis.text.y = element_text(), 
        axis.ticks = element_blank(), 
        plot.caption = element_text(hjust = 0, color = "grey25"),
        panel.background = element_blank()
        # plot.title.position = "plot"
  ) + 
  guides(fill = guide_legend(nrow = 1, reverse = TRUE, 
                             # title.position = "top",
                             # hjust = 0.5 #centers the title horizontally
                             # title.hjust = 0.5,
                             label.position = "top"))


# library(magick)
# -- Add Leauge logo to plot

ggsave(filename = "20241214 NHL Teams game state.png", 
       plot = last_plot(), 
       width = 6.5, height = 9, dpi = "retina")

# library(glue)
# plot_with_logo <- add_logo(plot_path = "20241211 NHL Teams game state.png",
#                            logo_path = glue("https://loodibee.com/wp-content/uploads/NHL-league-logo.png"),
#                            logo_position = "bottom right",
#                            logo_scale = 1000)