[Code]

---
title: "Netflix Text Analysis"
author: "NearAndDistant"
date: "11/07/2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r importing}
netflix_titles <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-04-20/netflix_titles.csv')
```

```{r cleaning / wrangling}

library(tidyverse)
library(lubridate)

netflix <- 
netflix_titles %>%
  ### eda1 : distribution of release over years
  mutate(
    date_added = parse_date(date_added , format = "%B %d, %Y"),
    month = as_factor(month(date_added)),
    year = as_factor(year(date_added))) %>%
  ### eda2 : string split and unnest for distribution of genre
  mutate(listed_in = strsplit(listed_in, ",")) %>%
  unnest(listed_in) %>%
  mutate(listed_in = trimws(listed_in)) %>%
  rename(genre = listed_in)

```

```{r}
library(tidytext)

words_unnested <- 
netflix %>%
  unnest_tokens("word" , description) %>%
  anti_join(stop_words , by = "word")

```

```{r}
library(widyr)
library(tidygraph)
library(ggraph)

showtext::showtext_auto()
sysfonts::font_add_google("Staatliches" , "staat")

netflix_logo_alpha <- magick::image_read("https://wallpapercave.com/wp/wp5063339.png") %>% 
                      magick::image_colorize(opacity = 20, color = 'white')

set.seed(42)

words_unnested %>%
  distinct(word , title , genre) %>%
  add_count(word , name = "word_total" , sort = TRUE) %>%
  mutate(word = str_to_title(word)) %>%
  filter(word_total >= 50) %>%
  widyr::pairwise_cor(word , title , sort = TRUE) %>%
  filter(correlation > 0.1) %>%
  igraph::graph_from_data_frame() %>%
  ggraph::ggraph(layout = "fr") +
  annotation_custom(grid::rasterGrob(netflix_logo_alpha, 
                               width = unit(1,"npc"), 
                               height = unit(1,"npc")), 
                    -Inf, Inf, -Inf, Inf) +
  ggraph::geom_edge_link(aes(alpha = correlation) , color = "white") +
  ggraph::geom_node_point(color = "white") +
  ggraph::geom_node_text(aes(label = name), repel = TRUE , color = "white" , size = 4) +
  labs(
    title = "Netflix Text Analysis",
    subtitle = "Which Words Chill Together Within The Netflix Catelogue Descriptions?",
    caption = "Source: kaggle.com | Graphic: @NearAndDistant") +
  theme(plot.title = element_text(family = "staat" , face = "bold" , color = colorspace::lighten("#E50914", 0.20) , size = 42, vjust = 6, hjust = -.05),
        plot.subtitle = element_text(family = "staat" , color = "white" , size = 16, vjust = 16 , hjust = -.06),
        plot.caption = element_text(family = "staat" , color = "white", size = 10 , vjust = -30 , hjust = -.04),
        panel.background = element_rect(fill = "grey20" , color = "grey20"),
        plot.background = element_rect(fill = "grey20" , color = "grey20"),
        legend.position = "none",
        plot.margin = margin(1,1.5,1.5,1.5, unit = "cm"))

```

```{r}
ggsave(here::here("Netflix.png"), dpi = 320, height = 10, width = 16)