---
title: "Exploration of Pixar Films Rankings"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Exploration of Pixar Films Rankings}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r setup}
library(pixarfilms)
library(ggplot2)
library(dplyr)
library(forcats)
library(gghighlight)
```


## Distribution of rankings

```{r}
pixar_rankings |>
  mutate(film = fct_reorder(film, ranking, .fun = median, .desc = TRUE)) |>
  ggplot(aes(x = film, y = ranking)) +
  geom_boxplot(fill = "#89B9F7") +
  coord_flip() +
  labs(
    title = "Distribution of Pixar Films rankings",
    subtitle = "Ordered by median ranking",
    x = "Film",
    y = "Ranking"
  ) +
  theme_bw()

# Save out for presentation use
# ggsave(
#   here::here("vignettes", "pixar_rankings.png"),
#   width = 8,
#   height = 5
# )
```


## Highlighting variation in rankings

A couple of things we notice from the boxplot above are that some films have a low variance in their rankings, while others have a higher variance. We can calculate the standard deviation of the rankings for each film to see which films have the lowest variance.

```{r}
summary_ranks <-
  pixar_rankings |>
  group_by(film) |>
  summarize(
    sd = sd(ranking),
    q1 = quantile(ranking, 0.25),
    q3 = quantile(ranking, 0.75),
    iqr = IQR(ranking),
    lower_bound = q1 - 1.5 * iqr,
    upper_bound = q3 + 1.5 * iqr
  ) |>
  arrange(sd)
print(summary_ranks)
```

We can see that `r head(summary_ranks$film)` have the lowest variance in their rankings. We can highlight these films in the boxplot above to see how they compare to the other films.

```{r}
pixar_rankings |>
  mutate(film = fct_reorder(film, ranking, .fun = median, .desc = TRUE)) |>
  ggplot(aes(x = film, y = ranking)) +
  geom_boxplot(fill = "#89B9F7") +
  coord_flip() +
  gghighlight(
    film %in% pull(head(arrange(summary_ranks, sd), 3), film),
    use_group_by = FALSE
  ) +
  labs(
    title = "Lowest variation of Pixar Films rankings",
    subtitle = "Ordered by median ranking",
    x = "Film",
    y = "Ranking"
  ) +
  theme_bw()

# Save out for presentation use
# ggsave(
#   here::here("vignettes", "pixar_rankings.png"),
#   width = 8,
#   height = 5
# )
```


We can also see that `r summary_ranks |> arrange(iqr) |> head(5) |> pull(film)` have the smallest IQR ranges in their rankings. We can highlight these films in the boxplot above to see how they compare to the other films.

```{r}
pixar_rankings |>
  mutate(film = fct_reorder(film, ranking, .fun = median, .desc = TRUE)) |>
  ggplot(aes(x = film, y = ranking)) +
  geom_boxplot(fill = "#89B9F7") +
  coord_flip() +
  gghighlight(
    film %in% pull(head(arrange(summary_ranks, iqr), 3), film),
    use_group_by = FALSE
  ) +
  labs(
    title = "Smallest IQRs of Pixar Films rankings",
    subtitle = "Ordered by median ranking",
    x = "Film",
    y = "Ranking"
  ) +
  theme_bw()

# Save out for presentation use
# ggsave(
#   here::here("vignettes", "pixar_rankings.png"),
#   width = 8,
#   height = 5
# )
```

Now let's highlight the highest variation in rankings. We can see that `r head(arrange(summary_ranks, desc(sd)), 5) |> pull(film)` have the highest variance in their rankings. We can highlight these films in the boxplot above to see how they compare to the other films.

```{r}
pixar_rankings |>
  mutate(film = fct_reorder(film, ranking, .fun = median, .desc = TRUE)) |>
  ggplot(aes(x = film, y = ranking)) +
  geom_boxplot(fill = "#89B9F7") +
  coord_flip() +
  gghighlight(
    film %in% pull(head(arrange(summary_ranks, desc(sd)), 3), film),
    use_group_by = FALSE
  ) +
  labs(
    title = "Highest variation of Pixar Films rankings",
    subtitle = "Ordered by median ranking",
    x = "Film",
    y = "Ranking"
  ) +
  theme_bw()

# Save out for presentation use
# ggsave(
#   here::here("vignettes", "pixar_rankings.png"),
#   width = 8,
#   height = 5
# )
```

Now let's highlight the outliers in the rankings.

```{r}
outlier_films <-
  pixar_rankings |>
  left_join(summary_ranks, by = "film") |>
  filter(ranking < lower_bound | ranking > upper_bound) |>
  distinct(film)

pixar_rankings |>
  mutate(film = fct_reorder(film, ranking, .fun = median, .desc = TRUE)) |>
  ggplot(aes(x = film, y = ranking)) +
  geom_boxplot(fill = "#89B9F7") +
  coord_flip() +
  gghighlight(
    film %in% outlier_films$film,
    use_group_by = FALSE
  ) +
  labs(
    title = "Pixar Films rankings with outliers",
    subtitle = "Ordered by median ranking",
    x = "Film",
    y = "Ranking"
  ) +
  theme_bw()

# Save out for presentation use
# ggsave(
#   here::here("vignettes", "pixar_rankings.png"),
#   width = 8,
#   height = 5
# )
```