Since the R language is not available for Heroku platform. Thus, I show my exploaratory analysis in this RPubs website. The Yelp review data are preprocessed and saved in the format of “.rds” for the sake of exploratory analysis. The dataset, Assoc, includes 2 sets of objects. One is the set of the most correlated words associated with “food” found in the review texts on the given restaurant, Chipotle Mexican Grill. The other is the corresponding correlations of those words. I would compare the datasets with 5, 3, and 1 star scores below.

The dataset shown below is selected with 5-star Yelp review.
The barplot of the most correlated words are shown below.

The dataset shown below is selected with 3-star Yelp review.
The barplot of the most correlated words are shown below.

The dataset shown below is selected with 1-star Yelp review.
The barplot of the most correlated words are shown below.

As you can see, for the higher star barplots, the words like “fresh”, “fast”, and “good” are shown in higher correlation. for lower star plot, no sentimental words are shown in the most 10 correlated word.