counting words in “lines” tokens

Question

I'm completely new in R, so this question may seem obvious. However, I didn't manage and didn't find solution

How can I count number of words within my tokens while they are lines (reviews, actually)? So, there is a dataset with reviews(reviewText) connected with ID of products(asin)

amazonr_tidy_sent = amazonr_tidy_sent%>%unnest_tokens(word, reviewText, token = "lines") amazonr_tidy_sent = amazonr_tidy_sent %>% anti_join(stop_words)%>%ungroup()

I tried to do in the following way

wordcounts <- amazonr_tidy_sent %>% group_by(word, asin)%>% summarize(word = n())

but it was not appropriate. I assume, that there is no way to count because line as a token cannot be "separated"

Thanks a lot

Answer 1

You can use unnest_tokens() more than once, if it is appropriate to your analysis.

First, you can use unnest_tokens() to get the lines that you want. Notice that I am adding a column to keep track of the id of each line; you could call that whatever you want, but the important thing is to have a column that will note which line you are on.

library(tidytext)
library(dplyr)
library(janeaustenr)


d <- data_frame(txt = prideprejudice)

d_lines <- d %>%
    unnest_tokens(line, txt, token = "lines") %>%
    mutate(id = row_number())

d_lines

#> # A tibble: 10,721 × 2
#>                                                                        line
#>                                                                       <chr>
#>  1                                                      pride and prejudice
#>  2                                                           by jane austen
#>  3                                                                chapter 1
#>  4  it is a truth universally acknowledged, that a single man in possession
#>  5                            of a good fortune, must be in want of a wife.
#>  6   however little known the feelings or views of such a man may be on his
#>  7 first entering a neighbourhood, this truth is so well fixed in the minds
#>  8 of the surrounding families, that he is considered the rightful property
#>  9                                 of some one or other of their daughters.
#> 10 "my dear mr. bennet," said his lady to him one day, "have you heard that
#> # ... with 10,711 more rows, and 1 more variables: id <int>

Now you can use unnest_tokens() again , but this time with words so that you will get a row for each word. Notice that you still know which line each word came from.

d_words <- d_lines %>%
    unnest_tokens(word, line, token = "words")

d_words
#> # A tibble: 122,204 × 2
#>       id      word
#>    <int>     <chr>
#>  1     1     pride
#>  2     1       and
#>  3     1 prejudice
#>  4     2        by
#>  5     2      jane
#>  6     2    austen
#>  7     3   chapter
#>  8     3         1
#>  9     4        it
#> 10     4        is
#> # ... with 122,194 more rows

Now you can do any kind of counting you want, for example, maybe you want to know how many words each line had in it?

d_words %>%
    count(id)

#> # A tibble: 10,715 × 2
#>       id     n
#>    <int> <int>
#>  1     1     3
#>  2     2     3
#>  3     3     2
#>  4     4    12
#>  5     5    11
#>  6     6    15
#>  7     7    13
#>  8     8    11
#>  9     9     8
#> 10    10    15
#> # ... with 10,705 more rows

Answer 2

By splitting each line using str_split we can count the number of words per line.

Some example data (containing newlines and stopwords):

library(dplyr)
library(tidytext)
d = data_frame(reviewText = c('1 2 3 4 5 able', '1 2\n3 4 5\n6\n7\n8\n9 10 above', '1!2', '1',
                          '!', '', '\n', '1', 'able able', 'above above', 'able', 'above'),
           asin = rep(letters, each = 2, length.out = length(reviewText)))

Counting the number of words:

by_line %>%
    group_by(asin) %>%
    summarize(word = sum(sapply(strsplit(word, '\\s'), length)))

   asin  word
  <chr> <int>
1     a    17
2     b     2
3     c     1
4     d     1
5     e     4

Note: in your original code most stopwords will not be removed because you split the data by line. Only lines that are exactly a single stopword will be removed.

To exclude stopwords from the wordcount use this:

by_line %>%
    group_by(asin) %>%
    summarize(word = word %>% strsplit('\\s') %>%
                  lapply(setdiff, y = stop_words$word) %>% sapply(length) %>% sum)

   asin  word
  <chr> <int>
1     a    15
2     b     2
3     c     1
4     d     1
5     e     0
6     f     0