Week 7: Working with HTML and JSON

Introduction

For this assignment, I used three books about data science in R to compare HTML and JSON. I entered the same book details into two files, then wrote R code to read each file into a data frame.

I wanted to see whether matching information would produce identical data frames. The comparison looks at both the book details and how R stores them.

Approach

I will select three books about data science in R and record the title, authors, year, publisher, and ISBN for each.

I will enter the same information into an HTML table and a JSON file by hand, then upload both files to GitHub. In R, I will use rvest to read the HTML table and jsonlite to read the JSON file into separate data frames.

I will compare their values, column names, and data types, then clean up any differences in structure and check whether the two data frames are identical.

The books

I chose these three books because they cover different parts of working with data in R, including preparing data, analyzing text, and selecting features for predictive models.

Title	Authors	Year	Publisher	ISBN
R for Data Science (2nd edition)	Hadley Wickham, Mine Çetinkaya-Rundel, Garrett Grolemund	2023	O’Reilly Media	978-1-492-09740-2
Text Mining with R: A Tidy Approach	Julia Silge, David Robinson	2017	O’Reilly Media	978-1-491-98165-8
Feature Engineering and Selection: A Practical Approach for Predictive Models	Max Kuhn, Kjell Johnson	2019	Chapman & Hall/CRC	978-1-138-07922-9

The two files

The code below reads books.html and books.json from the week-07 folder in my DATA607 repository.

library(tidyverse)
library(rvest)
library(jsonlite)

base_url <- "https://raw.githubusercontent.com/NawelMe/DATA607-Fall-2026/main/week-07"
html_url <- paste0(base_url, "/books.html")
json_url <- paste0(base_url, "/books.json")

HTML

This chunk displays the contents of the HTML file.

read_lines(html_url) |> cat(sep = "\n")

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<title>Data Science Books</title>
</head>
<body>
<table>
<thead>
<tr><th>Title</th><th>Authors</th><th>Year</th><th>Publisher</th><th>ISBN</th></tr>
</thead>
<tbody>
<tr><td>R for Data Science (2nd edition)</td><td>Hadley Wickham; Mine Çetinkaya-Rundel; Garrett Grolemund</td><td>2023</td><td>O'Reilly Media</td><td>978-1-492-09740-2</td></tr>
<tr><td>Text Mining with R: A Tidy Approach</td><td>Julia Silge; David Robinson</td><td>2017</td><td>O'Reilly Media</td><td>978-1-491-98165-8</td></tr>
<tr><td>Feature Engineering and Selection: A Practical Approach for Predictive Models</td><td>Max Kuhn; Kjell Johnson</td><td>2019</td><td>Chapman &amp; Hall/CRC</td><td>978-1-138-07922-9</td></tr>
</tbody>
</table>
</body>
</html>

I used <tr> for each row, <th> for the column headings, and <td> for the book details.

For books with multiple authors, I placed the names in one cell and separated them with semicolons. This keeps each book on one row, although the author names are stored as one text value.

I wrote the publisher Chapman & Hall/CRC as Chapman & Hall/CRC in the HTML source. The parser reads & as an ampersand.

JSON

This chunk displays the JSON file.

read_lines(json_url) |> cat(sep = "\n")

[
  {
    "title": "R for Data Science (2nd edition)",
    "authors": [
      "Hadley Wickham",
      "Mine Çetinkaya-Rundel",
      "Garrett Grolemund"
    ],
    "year": 2023,
    "publisher": "O'Reilly Media",
    "isbn": "978-1-492-09740-2"
  },
  {
    "title": "Text Mining with R: A Tidy Approach",
    "authors": [
      "Julia Silge",
      "David Robinson"
    ],
    "year": 2017,
    "publisher": "O'Reilly Media",
    "isbn": "978-1-491-98165-8"
  },
  {
    "title": "Feature Engineering and Selection: A Practical Approach for Predictive Models",
    "authors": [
      "Max Kuhn",
      "Kjell Johnson"
    ],
    "year": 2019,
    "publisher": "Chapman & Hall/CRC",
    "isbn": "978-1-138-07922-9"
  }
]

Each book is a JSON object with the same five fields. Unlike the HTML version, the JSON file stores the authors in an array.

For example, the authors of Text Mining with R appear as ["Julia Silge", "David Robinson"]. The year is written as a number, while the title, publisher, and ISBN are text.

Step 1: Read the HTML table

books_html <- read_html(html_url, encoding = "UTF-8") |>
  html_element("table") |>
  html_table()

books_html

# A tibble: 3 × 5
  Title                                            Authors  Year Publisher ISBN 
  <chr>                                            <chr>   <int> <chr>     <chr>
1 R for Data Science (2nd edition)                 Hadley…  2023 O'Reilly… 978-…
2 Text Mining with R: A Tidy Approach              Julia …  2017 O'Reilly… 978-…
3 Feature Engineering and Selection: A Practical … Max Ku…  2019 Chapman … 978-…

read_html() reads the page, html_element("table") selects its first table, and html_table() converts the table into a tibble.

I specified UTF-8 encoding to preserve the Ç in Çetinkaya-Rundel. An encoding mismatch would affect the comparison even if the names looked similar at first glance.

Step 2: Read the JSON file

books_json <- fromJSON(json_url)

books_json

                                                                          title
1                                              R for Data Science (2nd edition)
2                                           Text Mining with R: A Tidy Approach
3 Feature Engineering and Selection: A Practical Approach for Predictive Models
                                                   authors year
1 Hadley Wickham, Mine Çetinkaya-Rundel, Garrett Grolemund 2023
2                              Julia Silge, David Robinson 2017
3                                  Max Kuhn, Kjell Johnson 2019
           publisher              isbn
1     O'Reilly Media 978-1-492-09740-2
2     O'Reilly Media 978-1-491-98165-8
3 Chapman & Hall/CRC 978-1-138-07922-9

fromJSON() converts the JSON objects into a data frame with one row per book.

The authors column is a list column. Each entry holds a vector of author names rather than the single text value used in the HTML table.

Step 3: Compare the original data frames

identical(books_html, books_json)

[1] FALSE

This checks whether the two objects match exactly, including their structure and data types.

The following table shows three differences to examine.

tibble(
  `What` = c("Column names", "Type of object", "Authors column"),
  `HTML frame` = c(
    paste(names(books_html), collapse = ", "),
    class(books_html)[1],
    class(books_html$Authors)
  ),
  `JSON frame` = c(
    paste(names(books_json), collapse = ", "),
    class(books_json)[1],
    class(books_json$authors)
  )
) |> knitr::kable()

What	HTML frame	JSON frame
Column names	Title, Authors, Year, Publisher, ISBN	title, authors, year, publisher, isbn
Type of object	tbl_df	data.frame
Authors column	character	list

The HTML headings use capital letters, such as Title and Authors, while the JSON fields use lowercase names.

The HTML table is returned as a tibble, while the JSON file produces a standard data frame. Their author columns also differ because HTML stores the names together as text and JSON keeps them in a list.

These differences explain why matching book details alone are not enough for identical() to return TRUE.

Step 4: Make the structures consistent

I used one function to apply the same column names, object class, year type, and column order to both data frames.

tidy_books <- function(df) {
  df |>
    as_tibble() |>
    rename_with(tolower) |>
    mutate(
      authors = map_chr(authors, \(a) paste(a, collapse = "; ")),
      year = as.integer(year)
    ) |>
    select(title, authors, year, publisher, isbn)
}

html_clean <- tidy_books(books_html)
json_clean <- tidy_books(books_json)

html_clean

# A tibble: 3 × 5
  title                                            authors  year publisher isbn 
  <chr>                                            <chr>   <int> <chr>     <chr>
1 R for Data Science (2nd edition)                 Hadley…  2023 O'Reilly… 978-…
2 Text Mining with R: A Tidy Approach              Julia …  2017 O'Reilly… 978-…
3 Feature Engineering and Selection: A Practical … Max Ku…  2019 Chapman … 978-…

For the JSON data, map_chr() joins each vector of author names into one string separated by semicolons. For the HTML data, each entry is already a string, so this step leaves it unchanged.

I also converted the year to an integer in both versions instead of relying on the types chosen during import.

Step 5: Compare the cleaned data frames

identical(html_clean, json_clean)

[1] TRUE

all.equal(html_clean, json_clean)

[1] TRUE

all(html_clean == json_clean)

[1] TRUE

stopifnot(nrow(html_clean) == 3, ncol(html_clean) == 5)
stopifnot(identical(html_clean, json_clean))

These checks compare the cleaned objects and their values. If the files contain the same details and use matching author separators, the three comparisons should return TRUE.

The final checks require three rows, five columns, and identical cleaned data frames. If any requirement fails, R stops the report so I know something needs attention.

What I learned

The main difference in this example is how the two formats store multiple authors. I used one text value per book in HTML and an array in JSON. To compare the imported data frames directly, I converted the author arrays into text.

That choice works for this assignment. For an analysis of individual authors, I would keep the JSON list column or separate the names into rows instead.

I also paid attention to data types. HTML table values are parsed from text, and html_table() attempts to infer their types. JSON distinguishes numbers from strings in the source file. Explicitly converting the year helped make the comparison consistent.

The ISBNs include hyphens, which help keep them as text during import. I would still treat ISBNs as identifiers rather than numbers because they are labels, not quantities to calculate with.

This example gave me a way to compare file formats using the same information. Before combining imported data, I would check the structure as well as the values.

Assignment Video

Watch the HTML and JSON video walkthrough

AI Use

Anthropic. (2026). Claude Sonnet 5.5 [Large language model]. https://claude.ai. Accessed October 10, 2026.

I used Claude to help plan the comparison, draft the R code and report, and review the book details.