Sentiment analysis of Amazon 500,000 fine food reviews

• In this project I look at the review helpfulness and interaction between top reviewers in the comnnunity.
• Study the impact of time on the review score
• Perform NLP analysis on the review sentiments
• Analyze word structures using unsupervised methods

Link to the blog post

[Here is a blog post on the study on medium]

Dependecies and packages:

• Python 3.x
• Numpy
• Pandas
• Scikit learn
• Keras (Tensor flow backend)
• tqdm
• seaborn
• matplotlib
• bokeh
• IPython
• itertools
• collections

Repository content:

• Jupter notebook file: amazon_finefood_review_project1.ipynb
• Image folder (images): containing figures in the notebook
• A txt file : stopword_short_long_mod.txt, containg list of stop words
• MIT License file

Source dataset

The zipped csv file used in this study can be downloaded from:

• Web data: Amazon Fine Foods reviews
• Kaggle