DATA 607: Working with HTML and JSON

Reading and Comparing Book Data in R

Objective

This assignment demonstrates how to manually represent the same book data in HTML and JSON, load both formats into R data frames, and test whether the results are identical.

Selected books

I selected three books on Data Science and Machine Learning. Both R for Data Science and An Introduction to Statistical Learning have multiple authors. I recorded six attributes for each book: title, authors, year, publisher, ISBN, and subject.

The HTML file represents the books as rows in a table. The JSON file represents them as an array of book objects. Each object has the same six fields.

Bibliographic sources:

• R for Data Science, 2nd Edition — O’Reilly
• An Introduction to Statistical Learning, 2nd Edition — Springer
• Hands-On Machine Learning, 3rd Edition — O’Reilly

1. Load required packages

# Install once if necessary:
# install.packages(c("rvest", "jsonlite", "knitr"))
library(rvest)
library(jsonlite)
library(knitr)

2. Specify the HTML and JSON file locations

The files can be read locally while developing the assignment. Before publishing, upload both books.html and books.json to a public GitHub repository, then replace the two paths below with their public raw.githubusercontent.com URLs and render the report again. This allows anyone to rerun the code without needing the local files.

html_source <- "books.html"
json_source <- "books.json"

# Example public URL format after uploading to GitHub:
# html_source <- "https://github.com/sarahcunyms2026-source/DATA607-Week7-HTML-JSON/blob/main/books.html"
# json_source <- "https://github.com/sarahcunyms2026-source/DATA607-Week7-HTML-JSON/blob/main/books.json"

3. Read the HTML table into an R data frame

read_html() parses the HTML document, html_element() selects the table identified as books, and html_table() extracts the table cells. The year column is converted to integer so that its data type matches the JSON data.

html_doc <- read_html(html_source)
html_books <- html_doc |>
  html_element("#books") |>
  html_table(convert = FALSE) |>
  as.data.frame(stringsAsFactors = FALSE)

html_books$year <- as.integer(html_books$year)

kable(html_books, caption = "Books loaded from HTML")

Books loaded from HTML

title	authors	year	publisher	isbn	subject
R for Data Science, 2nd Edition	Hadley Wickham; Mine Çetinkaya-Rundel; Garrett Grolemund	2023	O’Reilly Media	9781492097396	Data Science and Machine Learning
An Introduction to Statistical Learning, 2nd Edition	Gareth James; Daniela Witten; Trevor Hastie; Robert Tibshirani	2021	Springer	9781071614174	Data Science and Machine Learning
Hands-On Machine Learning, 3rd Edition	Aurélien Géron	2022	O’Reilly Media	9781098125967	Data Science and Machine Learning

4. Read the JSON file into a separate R data frame

fromJSON() reads the JSON array and creates tabular data because every book object uses the same field names.

json_books <- fromJSON(json_source)
json_books <- as.data.frame(json_books, stringsAsFactors = FALSE)

kable(json_books, caption = "Books loaded from JSON")

Books loaded from JSON

title	authors	year	publisher	isbn	subject
R for Data Science, 2nd Edition	Hadley Wickham; Mine Çetinkaya-Rundel; Garrett Grolemund	2023	O’Reilly Media	9781492097396	Data Science and Machine Learning
An Introduction to Statistical Learning, 2nd Edition	Gareth James; Daniela Witten; Trevor Hastie; Robert Tibshirani	2021	Springer	9781071614174	Data Science and Machine Learning
Hands-On Machine Learning, 3rd Edition	Aurélien Géron	2022	O’Reilly Media	9781098125967	Data Science and Machine Learning

5. Compare both data frames

identical() checks whether both objects have the same structure and values, including column types, row order, and column order. ISBN values are stored as character strings so their digits are preserved.

comparison <- identical(html_books, json_books)
cat("Are the HTML and JSON data frames identical?", comparison, "\n")

Are the HTML and JSON data frames identical? TRUE 

# Fail clearly if the two data sources ever stop matching.
stopifnot(comparison)

Conclusion

The same three books were encoded by hand in two different file formats. HTML uses table elements (<table>, <tr>, and <td>), while JSON uses an array ([]) of objects ({}). R can import both formats using rvest and jsonlite. The final identical() check verifies that both sources produce equivalent R data frames, provided their values, order, and data types match. This exercise builds a foundation for future web scraping and API assignments.