Data 607 - Assignment 7

Instructions

Load libary

library(DT)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(htmltools)
library(readr)
library(aws.s3)
library(jsonlite)
library(RCurl)
library(XML)

Read URLs from GitHub

xml_url <- "https://raw.githubusercontent.com/Meccamarshall/Data607/main/Assignment%207/books.xml"
json_url <- "https://raw.githubusercontent.com/Meccamarshall/Data607/main/Assignment%207/books.JSON"
html_url <- "https://raw.githubusercontent.com/Meccamarshall/Data607/main/Assignment%207/books.html"

HTML

library(rvest)

## 
## Attaching package: 'rvest'

## The following object is masked from 'package:readr':
## 
##     guess_encoding

html_file <- getURL(html_url)
html_df <- readHTMLTable(html_file, which = 1)
html_df

##                            Title                          Author       ISBN-13
## 1 Black Girls Must Die Exhausted                     Jayne Allen 9780063137905
## 2        The Coldest Winter Ever                  Sister Souljah 9780743270106
## 3           One of the Good Ones Malika Moulite, Martiza Moulite 9781335426253
##   Publication Year Pages
## 1             2021   368
## 2             2005   368
## 3             2022   352

datatable(html_df)

JSON

json_df <- as.data.frame(fromJSON(json_url))
names(json_df) <- c("Title","Author", "ISBN-13", "Publication Year","Pages")
json_df

##                            Title                          Author       ISBN-13
## 1 Black Girls Must Die Exhausted                     Jayne Allen 9780063137905
## 2        The Coldest Winter Ever                  Sister Souljah 9780743270106
## 3           One of the Good Ones Malika Moulite, Martiza Moulite 9781335426253
##   Publication Year Pages
## 1             2021   368
## 2             2005   368
## 3             2022   352

datatable(json_df)

XML

xml_file <- getURL(xml_url)
xml_df <- xmlToDataFrame(xml_file)
xml_df

##                            title                       authors        isbn13
## 1 Black Girls Must Die Exhausted                   Jayne Allen 9780063137905
## 2        The Coldest Winter Ever                Sister Souljah 9780743270106
## 3            One of the Good One Malika MouliteMartiza Moulite 9781335426253
##   publicationyear pages
## 1            2021   368
## 2            2005   368
## 3            2022   352

datatable(xml_df)