textmining-subtitles.Rmd

---
title: "textmining subtitles"
output: html_notebook
---

[source material](https://joenoonan.se/post/cats/)

## Use movie subtitles for textmining

Movie subtitles are stored in `.srt` files, find a movie or a play that has a srt file and do analysis.

```{r libraries, warning=FALSE, message=FALSE}
library(tidyverse)
# install.packages('srt')
library(srt) # reads subtitle text
library(tidytext)
library(scales)
library(paletteer)
library(showtext)
library(ggtext)

font_add_google("Lato","Lato")
showtext_auto()

cats = srt::read_srt("./Cats_Netflix_en.srt")
head(cats)
```


## clean text 

Use tidytext library to clean the text by using the `stopwords` and to count the words

```{r}
cats_clean_count = cats %>% 
  unnest_tokens(word, subtitle) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = T)

head(cats_clean_count)
```


stem the words for proper text analysis, looking for only word cat stems

```{r}
cats_clean_count %>% 
  filter( str_detect(word, "\\bcat\\b|\\bcats\\b|\\bcat's\\b"))
```

look for when the word cat is said 

```{r}
cat_pattern = "\\bcat\\b|\\bcats\\b|\\bcat's\\b"

total_cat_mentions = cats %>% 
  rename(time = start) %>% 
  select(time, subtitle) %>% 
  mutate(cat_mentions = str_count(subtitle, cat_pattern),
         cumulative_cat_mentions = cumsum(cat_mentions))

head(total_cat_mentions)
```




```{r}


total_cat_mentions %>% 
  ggplot(
    aes(x= time,
        y=  cumulative_cat_mentions,
        col= cumulative_cat_mentions
        )
  )+
  geom_line(show.legend = F, size= 1.5)+
  scale_x_time() +
  ggdark::dark_mode()+
  scale_color_paletteer_c(`"gameofthrones::margaery"`) +
  labs(
    title = "\nCumulative mentions of `cats` in Cats (2019)",
    subtitle = "words used from subtitle text",
    x= "\nFilm time",
    # y= "Cumulative mentions of 'cats'\n",
    caption = "\n@unicornCoder | Nov 26, 2022 | Source: @JoeDNoonan"
  )+
  theme(
    text = element_text(family = "Lato"),
    plot.title = element_text(hjust = 0.5, size = 13, color= '#ffcc66',face = 'bold'),
    plot.subtitle = element_text(hjust = 0.5, color = 'grey60'),
    plot.caption = element_text(hjust = 0, size = 11, color = 'grey50'),
    axis.text.y = element_text(size = 11, color= '#ffcc66'),
    axis.title.x = element_text(size = 11, color= 'grey70'),
    
    axis.title.y = element_blank()
  )

```