/* ---- Google Analytics Code Below */

Tuesday, March 27, 2018

Data Mining Text Reviews

An example of analyzin reviews using analytical methods.

Tidy Text Mining Beer Reviews

On R, tidytext, clustering, classification, correlation, tfidf, knn, dataviz, web scraping

Craft beer is a huge market. Beer reviews from fellow beer drinkers help customers navigate it.

From 2013 to 2016, the craft beer industry nearly doubled in value from $14.3 to $23.5 billion. The number of craft breweries jumped from 2,768 to 5,005, and production ramped up from 15 to 25 million barrels a year.

BeerAdvocate was founded in 1996 as a forum to review, rate and share ideas about beer. Today, the site houses millions of reviews for hundreds of thousands of beers, both domestic and international.

Users can rate beers and optionally provide text explaining their assessment. BeerAdvocate takes a thorough approach to beer evaluation, asking users to rate each on several features: look, smell, taste, feel, and overall. The platform also provides measures to indicate the spread or consensus of reviews, both in the rDev (a single rating's deviation from the overall mean rating) and the pDev (deviation within a beer's ratings).   ... "

No comments: