I recently learned to do text analytics formally from my professor, Prof. Chris Maurer at UVA, and thought that it would be fun to do a practice of frequency analysis on reviews about my favorite restaurant in Las Vegas, Momofuku.
I basically:
1. obtained JSON files (consisting of business and review information) from Yelp using Yelp API
2. found the information related to the restaurant I am interested in (Momofuku Las Vegas) and collated it into a CSV file. This dataset consists of reviews from Jan 2017 till Dec 2019, and has 2373 reviews.
3. used the code given by Prof. Maurer and from Text Mining with R website (https://www.tidytextmining.com/tfidf.html) to read the CSV file, clean the text data, perform frequency analysis, and make visualizations.
Some interesting findings:
This Momofuku had the most number of reviews in July 2018.
I am surprised to find that the average monthly review rating of this Momofuku in Vegas is a clean 4.0. Hence, I decided to show the changes in review count over time instead. There is an upward trend for the review count from April 2018 till July 2018. I am not exactly sure why the increase, but it could be simply because more people go to Vegas on Independence Day (as shown in the smaller spike in review count in July 2017 as well). Or I would like to associate it with the decease of Anthony Bourdain in June 2018, and David Chang (as a good friend of Anthony Bourdain)’s Netflix show Ugly Delicious kind of reminds people about Anthony Bourdain’s Parts Unknown. Hence, people decided to visit this restaurant. (a bit of a stretch but oh well).
“ramen,” “pork,” and “bun” are the most frequent words in the reviews.
These words (and “food”) appears in 1000–1325 reviews (df value range is 1000–1325), having appearances in ~50% of the reviews.
When I did bigram frequency analysis, these common words appear to mean “pork belly,” “pork ramen,” “belly buns.” Another common bigram is “David Chang,” which represents the owner of the Momofuku chain and is a celebrity chef.
What do these mean? People go to Momofuku for its famous pork ramen and pork belly buns. I mean, the ramen is delicious.
Not only that, customers who went to Momofuku for the ramen and buns, were willing to write good reviews about the restaurant. David Chang has also established a good brand name for himself, attracting people (like me)to go to his restaurants.
Milk Bar and Momofuku
Milk Bar, located right next to Momofuku, is also a topic of discussion among Yelpers. Its famous cereal milk products (and crack pie) seem to be tasty desserts after a bowl of pork ramen. This bigram frequency graph just shows that the most commonly discussed topics in the reviews are mostly food (spicy cucumber, scallion noodle, bossam, etc.). David Chang’s show Ugly Delicious is also mentioned frequently, again showcasing his brand power. Customer service, the need to wait, and dining experience are also discussed often. Based on the average score, the reviewers are not unhappy about these aspects.
Extras
Most of my time for this practice was spent on getting the data from API and transforming them into CSV. I also played around with the color palettes for visualization. It is a straightforward practice since the codes for frequency analysis are already available for me. I just thought it’s fun to learn something about Momofuku.
And I low key miss Las Vegas (and Los Angeles). And this is a picture of Bossam (not pork but duck). It is the reason why I visited Momofuku during every Vegas trip.
Comments