The document describes a text analysis project that involved:
1. Crawling reviews for the Moto G (2nd gen) from Flipkart.com.
2. Creating a term document matrix and word cloud to analyze terms.
3. Using latent semantic analysis for dimension reduction.
4. Clustering reviews based on terms and documents.
5. Analyzing ratings and comparing sentiments in reviews to ratings.
1. TEXT ANALYTICS
Analysis of reviews fetched
from FLIPKART.COM for
MOTO-G (2nd gen)
3/19/2015 Roma Agrawal (A14026) 1
2. Objective
• Web Crawling from www.filpkart.com for Mobile “MOTO G (2nd gen)”
• Creation of Term Document Matrix and WordCloud
• Dimension Reduction using Latent Semantic Analysis
• Clustering on the basis of both Terms and Documents
• Analysis of ratings given
• Comparison of sentiments expressed in reviews and ratings given
3/19/2015 Roma Agrawal (A14026) 2
15. Analysis of ratings given
3/19/2015 Roma Agrawal (A14026) 15
No Rating was missing
16. What review text is saying!
3/19/2015 Roma Agrawal (A14026) 16
Some agreement Between
“what review text is saying” and
“what rating given” is shown by
this count
Not able to create polarity for 6
reviews, replaced by group
average polarity
17. 3/19/2015 Roma Agrawal (A14026) 17
Classification on “satisfaction”
Top 6 imp words which have
negative meaning
Top 6 imp words which have
positive meaning