Evaluating Yelp reviews

We read in the processed Yelp reviews using this script and print out some statistics of the data:

reviews <- read.csv("c:/Users/Dan/yelp_academic_dataset_review.csv")

I usually take a look at some of the data once loaded to visually check that things are working as expected. We can do this with a head() function call:

head(reviews)

Summary data

All of the columns appear to be correctly loading. Now, we can look at summary statistics for the data:

summary(reviews)

There are several points in the summary worth noting:

• Some of the data points I had assumed would be just TRUE/FALSE, 0/1 have ranges instead; for example, funny has a max value over 600; useful has a max 1100, cool has 500.
• All of the IDs (users, businesses) have been mangled. We could use the user file and the business file to come up with exact references.
• Star ratings are 1-5, as expected. However, the mean and median are about a 4, which I take as many people only take the time to write good reviews.

Review...