Text Analytics Using R - Part C: Extraction of Samsung S4 white reviews from Amazon.com and Amazon. India

Hi Folks!!

 Welcome to the 3rd part of the series where we would be extracting the reviewers name , date , rating and review for the product galaxy s4

 The given below is the code to extract for the .com and .in and combining them into the new datafile

 Code:
--library(RCurl)
--library(XML)
--library(rvest)
--init <- "http://www.amazon.in/Samsung-Galaxy-S4-GT-I9500-White/product  reviews/B00CL4HXQC"
--crawlCandicate = "ref=cm_cr_pr_btm_link_"
--base <- "http://www.amazon.in"
--num <- 3
--doclist <- list()
--anchorlist <- vector()
--j <- 0
--while(j<num){
--if(j==0){
--doclist[j+1] <- getURL(init)
--} else{
--doclist[j+1] <- getURL(paste(base,anchorlist[j+1],sep = ""))
--}
 -- doc <- htmlParse(doclist[[j+1]])
  --anchor <- getNodeSet(doc,"//a") # capture all the 'a' tags which contains all the 
  --anchor <- sapply(anchor,function(x) xmlGetAttr(x,"href")) 
--anchor <- anchor[grep(crawlCandicate,anchor)] 
 -- anchorlist <- c(anchorlist,anchor)
  --anchorlist <- unique(anchorlist)
  --j <- j + 1
--}

 --reviews <- c() 
--ratings=c()
--reviewers = c()
--date_of_review = c()
--for (i in 1:3)
--{
  --doc <- htmlParse(doclist[[i]])
  --l = getNodeSet(doc,"//div/span[@class='a-size-base review-text']")
  --l1 <- sapply(l,xmlValue)
  --rateNodes=getNodeSet(doc,"//div/a/i/span[@class='a-icon-alt']")
 -- rates=sapply(rateNodes,xmlValue)
  --reviewerNodes = getNodeSet(doc,"//div/span[@class='a-size-base a-color-secondary reviewbyline']")
  --reviewer=sapply(reviewerNodes,xmlValue)
  --dateNodes = getNodeSet(doc,"//div/span[@class='a-size-base a-color-secondary review-date']")
  --dates=sapply(dateNodes,xmlValue)
  --ratings=c(ratings,rates)
  --reviews <- c(reviews,l1)
  --reviewers <- c(reviewers,reviewer)
  --date_of_review <- c(date_of_review,dates)
--}

 --reviews.data <- data.frame(reviews,site = "Amazon.in")
--ratings.data <- data.frame(ratings,site = "Amazon.in")
--reviewers.data <- data.frame(reviewers,site = "Amazon.in")
--date_of_review.data <- data.frame(date_of_review,site = "Amazon.in")
--Amazon_IN <- data.frame(reviews = reviews.data[1:30,],
                        ratings = ratings.data[1:30,-2],
                        reviewers = reviewers.data[1:30,-2],
                        date_of_review = date_of_review.data[1:30,-2])
#Amazon.com reviews
--library(RCurl)
--library(XML)
--library(rvest)
--init <- "http://www.amazon.in/Samsung-Galaxy-S4-GT-I9500-White/product-reviews/B00CL4HXQC"
--crawlCandicate = "pageNumber="
--base <- "http://www.amazon.com"
--num <- 3
--doclist <- list()
--anchorlist <- vector()
--j <- 0
--while(j<num){
  --if(j==0){
    --doclist[j+1] <- getURL(init)
    --} else{
    --doclist[j+1] <- getURL(paste(base,anchorlist[j+1],sep = ""))
    --}
 -- doc <- htmlParse(doclist[[j+1]])
  --anchor <- getNodeSet(doc,"//a") 
  --anchor <- sapply(anchor,function(x) xmlGetAttr(x,"href")) 
  --anchor <- anchor[grep(crawlCandicate,anchor)
  --anchorlist <- c(anchorlist,anchor)
  --anchorlist <- unique(anchorlist)
  --j <- j + 1
--}
--reviews <- c() 
--ratings=c()
--reviewers = c()
--date_of_review = c()
--for (i in 1:3)
--{
 --doc <- htmlParse(doclist[[i]])
--l = getNodeSet(doc,"//div/span[@class='a-size-base review-text']")
-- l1 <- sapply(l,xmlValue)
 --rateNodes=getNodeSet(doc,"//div/a/i/span[@class='a-icon-alt']")
 --rates=sapply(rateNodes,xmlValue)
 --reviewerNodes = getNodeSet(doc,"//div/span[@class='a-size-base a-color-secondary review-byline']")
 -- reviewer=sapply(reviewerNodes,xmlValue)
  --dateNodes = getNodeSet(doc,"//div/span[@class='a-size-base a-color-secondary review-date']")
  --dates=sapply(dateNodes,xmlValue)
  --ratings=c(ratings,rates)
  --reviews <- c(reviews,l1)
  --reviewers <- c(reviewers,reviewer)
  --date_of_review <- c(date_of_review,dates)
--}reviews.data <- data.frame(reviews,site = "Amazon.com")
--ratings.data <- data.frame(ratings,site = "Amazon.com")
--reviewers.data <- data.frame(reviewers,site = "Amazon.com")
--date_of_review.data <- data.frame(date_of_review,site = "Amazon.com")
--Amazon_COM <- data.frame(reviews = reviews.data[1:30,],
                         ratings = ratings.data[1:30,-2],
                         reviewers = reviewers.data[1:30,-2],
                         date_of_review = date_of_review.data[1:30,-2])
--Amazon_ALL=rbind(Amazon_IN,Amazon_COM)

 Explanation:Majority of the code has been explained in Part A and Part B of the blog. Here we have attached the code for the amazon.in and amazon.com where we can get the respective reviews









Comments

Post a Comment

Popular posts from this blog

SAS The language of Analytics Blog 1: Working with simple data sets

Image
Welcome back to the world of analytics. Today's scope of coverage is how to write a simple SAS Programme.   About SAS:  This is a very powerful tool that has been used across companies to take advantage of its statistical   capability.  The methodology adopted in this blog is through questions and understanding the output and code lines: You have a text file called scores.txt containing information on gender (M or F) and four test scores (English, history, math, and science). Each data value is separated from the others by one or more blanks.   a.     Write a DATA step to read in these values. Choose your own variable names. Be sure that the value for Gender is stored in 1 byte and that the four test scores are numeric. b.     Include an assignment statement computing the average of the four test scores. c.     Write the appropriate PROC PRINT statements to list the...
Read more

Kabaddi Match: Lets meet at the arena!! Aa jao Dam Dikhane!!!

Image
Kabaddi is one of the oldest sports of India.   It has been the national game of Bangladesh and Nepal. The game is a test of strength and unity. The game continues till the other team members oust the last player. The game tests ones mental, physical and interpersonal skills. The onus of one lies on other members hence the trust is the other key factor. The statistical programming languages are on the same fighting arena. Lets look around us we have R, Python, SPSS, SAS, Stata among many other. Which one is the best? 1)We have problem statement on which analysis needs to be done any language can be used to build up on the solution as the languages are a tool whereas logic remains same behind these. 2)R is a free open courseware whereas others are costly in terms of licensing costs. Meanwhile R has been used most freely SAS/Stata is more powerful to handle more statistical languages and techniques. 3)R being free is the most documented and used for the research and t...
Read more

Replace your Social Media with the New Age Social Media: Fitness with Motivation

Image
I had quit Social Media a year ago especially Instagram and as any other millennial hooked to Social Media the minute I woke up I wanted to get hold my Phone and checking immediately the feeds. More than the feeds it was just adding to the ever growing anxiety and also stealing away the beautiful moments which you spend online. I took a very drastic step some 7-8 Months Back to quit Instagram. It was not sudden but a gradual deaddiction mechanism I followed. I deactivated Instagram and we are allowed to activate/deactivate once per week. It was kind of sheer stupidity of my actions that I was following then came the Day I finally deleted my Instagram Account. Post that day It was a sudden vacuum that came but what really helped me was I suddenly have the new found time to pursue my Activities. I started drawing and below is one of my new passtimes I also started actively working out when I came across an interesting App Strava. I was suddenly making new friends and...
Read more