- Implemented a web scraper to fetch live hotel reviews from TripAdvisor website.
- Conducted data pre-processing and cleaning by removing stop words, punctuation, special characters, numbers, and white-spaces from reviews.
- Performed tokenization and stemming of reviews, and built a corpus out of it.
- Calculated the sentiment score of each review by comparing it's tokens with positive and negative lexicon and the overall sentiment score of the hotel depending on the number of positive and negative reviews it received.
- Visualized the results using bar-plot and box-plot charts.
- Sentiment score of first hundred reviews
- Positive and negative words in a single review
- Positive word count in first hundred reviews
- Negative word count in first hunderd reviews
- Overall sentiment score of JW Marriot hotel
R
RStudio