Tripadvisor NLP Analysis

A Sentiment Analysis and Wordcloud Vizualization on Restaurant Reviews

Define the subject

Analyze customer reviews on Tripadvisor Restaurants section to improve customer experience when choosing a restaurant.

Define business case

When choosing a restaurant, one can easily use stars and rankings to filter only those restaurants which have good ratings. But other people's feedbacks will most often make you decide whether you want to book a place or not. How can NLP help retrieve qualitative information provided by customer reviews?

Define a plan

Scrape Tripadvisor customer reviews. Clean and preprocess the data. Get a grasp of dataset with EDA. Process reviews to extract clean, trimmed text. Score reviews with sentiment analysis. Visualize wordcloud of review keywords.

Data collection

Webscraping data with Octoparse.

Data preparation

Clean data. Plot EDA results. Tokenize textual data. Filter stop words. Clean and lemmatize text.

Data modeling

Process sentiment analysis. Vectorize words and bigrams counts. Visualize polarized word clouds.

Deliverables

• Github
• Trello
• Slides
• Web-view Slides
• Data Prepping on colab
• Data Modeling on colab

Useful links

• https://www.analyticsvidhya.com/blog/2020/04/beginners-guide-exploratory-data-analysis-text-data/
• https://towardsdatascience.com/a-beginners-guide-to-sentiment-analysis-in-python-95e354ea84f6
• https://towardsdatascience.com/generate-meaningful-word-clouds-in-python-5b85f5668eeb
• https://youtu.be/_xP5Tybr6m4

Documentation

• https://www.octoparse.com/
• https://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.CountVectorizer.html
• https://textblob.readthedocs.io/en/dev/quickstart.html
• https://amueller.github.io/word_cloud/