Currently I’m working on an interdisciplinary research topic of Amazon Product Reviews. This page briefly describes my research experience and interests.
Amazon Product Reviews
Amazon Product Reviews is an interdisciplinary research area which aims to understand the overall product reviews and generate key insights from it, specifically highlighting the disfunctional products. Based on the analysis, we can further try to analyze the imapct of these products on the customers buying it. However we are also planning to develop a platform where we can share these findings so that the customers may get benefitted from the findings and thus, it can help them to decide upon the purchase of the flagged product.
I am working on this research advised by Prof. David Goldberg AI Research Lab @ SDSU.
Some of the questions I am trying to answer at present:
- How can we use the customer reviews to get insights upon disfunctional products?
- How do we identify the underlying data patterns between all the flagged products using these reviews?
- How do we understand & quantify the associated flagged products which are related to our discfunctional product analysis?
- How do we create a way for highlighting these products and inform customers, which will inturn affect their product purchasing decision?
Discourse Analytics
Discourse Analytics is an approach to understanding the nature and function of world-level to root-level discourse using digital data sources like speech transcripts, debates, published articles, blogs, tweets, and other forms of documentation to derive automated inferences, conclusions, predicting impact, etc. I am specifically interested in conducting the discourse studies using NLP approaches like text analytics, topic discovery / topic modeling, semantic analysis, text embedding technologies to better embed language, etc.
Some of the questions I tried to examine:
- How do we use language technologies to trace and identify social issues from speech transcripts?
- How do we extract and interpret latent ideologies, topics/shades from large unstructured corpora?
- How can a study like this be made useful for analysts, strategists and the people to assess multiple candidates and their attitude & discourse?
New Interests In
- Co-speech Gesture Analysis from Text Data
- Analyzing Discourse from 3 Levels
- What gets delivered? (Speech)
- What gets reported? (News Reports)
- What gets discussed? (Social Media)