What if we looked at one of the top-rated firms to see where the ratings spread is? We could use the following script:

# range of ratings for business with most ratings
library(sqldf)
most_rated = sqldf("select * from reviews where business_id = '4JNXUYY8wbaaDmk3BPzlWw' ")
hist(most_rated$stars)  

The script takes one of the top rated IDs, accesses all of their ratings, and displays a histogram of the same:

So, for one of the top rated firms there are very few low ratings.