hockey tidytuesday

library(gghighlight)
library(tidyverse)

# load data
raw <- read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2020/2020-03-03/game_goals.csv')

# 8 best scorers (there are 8 colors in scale_color_brewer Dark2)
best <- raw %>% 
        group_by(player) %>%
        summarize(goals = sum(goals)) %>%
        arrange(-goals) %>%
        pull(player) %>%
        head(8)

# cumsum
dat <- raw %>% 
       filter(player %in% best) %>%
       arrange(player, date) %>%
       group_by(player) %>% 
       mutate(`Goals scored` = cumsum(goals),
              `Games played` = 1:n())

# plot
p <- ggplot(dat, aes(`Games played`, `Goals scored`, color = player, group = player)) +
    geom_line(alpha = 0) + # hack: gghighlight needs a line
    geom_smooth(se = FALSE) +
    xlab('Games played') +
    ylab('Goals scored') +
    gghighlight(use_direct_label = TRUE, 
                label_params = list(size = 2.5, fill = 'white',
                                    box.padding = .1,
                                    label.size = .01,
                                    alpha = .9)) +
    theme_minimal() +
    scale_colour_brewer(palette = "Dark2") +
    theme(panel.grid = element_blank())
ggsave(p, file = '~/Downloads/tidytuesday.png', width = 8, height = 6)