Sentiment Analysis Using Twitter

  • Published on
    11-Feb-2017

  • View
    24

  • Download
    12

Transcript

Sentiment Analysis

Sentiment Analysis Using TwitterNatural language Processing

VS

Sentiment Comparison between the 2 Giant companies

Objective The objective of this sentiment analysis is to understand customers review/opinion for both the organizations. To understand which company is more liked/loved by the customers.

First we have extracted tweeter data in R platform(around 2500 tweets).

Next we cleaned the data and gave it a structural form for our analysis.

Then, we made 3 world clouds, to understand the frequencies of each kind of words: 1) for the whole data, 2) For Positive response in favor of the organization. 3) Negative response of the organization.

Next we counted the occurrence of word and plotted a histogram to understand the distribution of words following its occurrence in twitter.

Then, we used SVM (machine learning algorithm to find out the accuracy)

Package used tweetR, stringr, ROAuth, plyr, ggplot2,tm, RcolorBrewer, httr, wordcloud, sentimentr, Rcurl, snowball, e1071

Methodology/Process

WordCloud of Flipkart and Amazon

Positive and Negative WordCloud of Flipkart andAmazon

Positive and Negative WordCloud of Flipkart andAmazon

Visualization using HistogramVisualization using Histogram

Interpretation of AnalysisThe mean Score of total word count of both the organization is positive, which indicates that there are positive reactions for both the organization.We have classified the response using SVM ( 0- positive, 1- negative) with 87.40% and 84.50%.

Interpretation of Analysis

8

ConclusionBoth the organization has positive customer response with positive mean value.But, Amazon has higher positive response rate when compared to Flipkart, which indicates that currently Amazon is ruling the market with higher sales and revenue/profit (as customers have a positive reviews for it).Conclusion

Recommended

View more >