Earn the Text Mining Certificate
Practice the exercises below. When you feel ready, attempt the quiz to earn a verifiable certificate you can share on LinkedIn.
Text Mining Exercises in R: 20 Practice Problems
Twenty practice problems on text mining in R: tokenization, word counts, tf-idf, n-grams, sentiment, stop-words. Hidden solutions.
RRun this once before any exercise
library(dplyr)
library(stringr)
library(tibble)
library(cld3)
library(quanteda)
library(textstem)
library(tidyr)
library(tidytext)
library(wordcloud)
Exercise 1: Tokenize a sentence
Difficulty: Beginner.
Show solution
RInteractive R
str_split("hello world from R", " ")[[1]]
Exercise 2: Lowercase tokens
Difficulty: Beginner.
Show solution
RInteractive R
str_to_lower(c("Hello","WORLD","R"))
Exercise 3: Word frequencies
Difficulty: Intermediate.
Show solution
RInteractive R
words <- str_split("the cat sat on the mat the dog ran", " ")[[1]]
table(words)
Exercise 4: Remove stopwords
Difficulty: Intermediate.
Show solution
RInteractive R
stops <- c("the","a","is","on","to","and")
words <- c("the","cat","is","on","the","mat")
words[!words %in% stops]
Exercise 5: tidytext unnest_tokens
Difficulty: Intermediate.
Show solution
RInteractive R
df <- tibble(id = 1, text = "Hello world from R")
tidytext::unnest_tokens(df, word, text)
Exercise 6: Sentiment with bing lexicon
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(id = 1:2, text = c("I love R", "This is terrible"))
df |>
tidytext::unnest_tokens(word, text) |>
inner_join(tidytext::get_sentiments("bing"), by = "word") |>
count(id, sentiment)
Exercise 7: TF-IDF
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(doc = c("d1","d2"), text = c("R is great R is powerful", "Python is great"))
df |>
tidytext::unnest_tokens(word, text) |>
count(doc, word) |>
tidytext::bind_tf_idf(word, doc, n)
Exercise 8: Bigrams
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(id = 1, text = "the cat sat on the mat")
tidytext::unnest_tokens(df, bigram, text, token = "ngrams", n = 2)
Exercise 9: Word cloud (concept)
Difficulty: Intermediate.
Show solution
RInteractive R
# wordcloud::wordcloud(words, freq, min.freq = 1)
Exercise 10: Document-term matrix
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(doc = c("d1","d2"), text = c("hello world", "world R"))
df |>
tidytext::unnest_tokens(word, text) |>
count(doc, word) |>
tidyr::pivot_wider(names_from = word, values_from = n, values_fill = 0)
Exercise 11: Text length
Difficulty: Beginner.
Show solution
RInteractive R
str_count(c("hello world", "hi"), "\\w+")
Exercise 12: Detect language (concept)
Difficulty: Advanced.
Show solution
RInteractive R
# cld3::detect_language("Bonjour le monde")
Exercise 13: Replace contractions
Difficulty: Intermediate.
Show solution
RInteractive R
str_replace_all("don't can't won't", c("don't" = "do not",
"can't" = "cannot",
"won't" = "will not"))
Exercise 14: Stem words
Difficulty: Advanced.
Show solution
RInteractive R
SnowballC::wordStem(c("running","runner","runs"))
Exercise 15: Frequent terms by group
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(group = c("A","A","B"), text = c("R is great", "R is powerful", "Python is also great"))
df |>
tidytext::unnest_tokens(word, text) |>
count(group, word, sort = TRUE) |>
group_by(group) |> slice_head(n = 3)
Exercise 16: Document similarity (cosine)
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(doc = c("d1","d2"), text = c("R is great", "R is great"))
dtm <- df |> tidytext::unnest_tokens(word, text) |>
count(doc, word) |>
tidyr::pivot_wider(names_from = word, values_from = n, values_fill = 0)
v1 <- as.numeric(dtm[1,-1]); v2 <- as.numeric(dtm[2,-1])
sum(v1*v2) / (sqrt(sum(v1^2)) * sqrt(sum(v2^2)))
Exercise 17: Top tf-idf per doc
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(doc = c("d1","d1","d2","d2"), word = c("r","stats","python","stats"), n = c(2,1,3,1))
df |> tidytext::bind_tf_idf(word, doc, n) |>
group_by(doc) |> slice_max(tf_idf, n = 2)
Exercise 18: Filter very rare/common words
Difficulty: Advanced.
Show solution
RInteractive R
df <- tibble(doc = c("d1","d1","d2"), word = c("r","stats","r"), n = c(2,1,3))
df |> group_by(word) |> filter(n() >= 2)
Exercise 19: Lemmatize (textstem)
Difficulty: Advanced.
Show solution
RInteractive R
# textstem::lemmatize_words(c("running","runs","ran"))
Exercise 20: Word context (kwic)
Difficulty: Advanced.
Show solution
RInteractive R
# quanteda::kwic(quanteda::tokens("the quick brown fox"), pattern = "quick", window = 2)
What to do next
- NLP-Exercises (coming), language modeling beyond bag-of-words.
- stringr-Exercises (shipped), string ops.
Ready to earn the Text Mining Certificate?
The quiz is concept-based and respects your time: pass it once and your verifiable certificate is yours to share on LinkedIn, your resume, or your portfolio. Take it when you feel comfortable with the material.
Attempt the quiz→