Data Analysis on Covid-19's Impact on the Restaurant Business

Posted on Aug 3, 2020
The skills I demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.


As the novel Coronavirus has spreading around the world, the collateral damage has been far and wide, with the restaurant industry suffering one of the hardest hits. According to Statista data, the number of seated diners in restaurants worldwide has declined by over 54.4% year over year since Feb – July 2020.

Given such a harrowing state of affairs for the restaurant industry, I wanted to take a deeper look into restaurants’ business operations during the pandemic across various cuisine types and locations, and understand which services and measures implemented has been helpful for restaurants.

For this analysis, I decided to use Yelp restaurants review rating before and after 3/1/2020 as a proxy for restaurants’ overall performance before and after the pandemic started, and delve into various information posted on each restaurants’ Yelp listing to gain insights into which features potentially contributed to their performance.

Data Collection and Transformation

I scraped the data from using Python Scrapy to gather information about restaurants across three major U.S. cities (New York City, San Francisco, and Austin, TX) and seven cuisine types (American, Italian, Mediterranean, Mexican, Japanese, Chinese, and Thai), which adds up to a total of 4800 restaurants.

The fields collected on each restaurant include price, category, number of total reviews, average rating, price range, business hours, the review rating and content since 3/1/2020, and the Covid-19 Updates details which include updated services such as delivery, takeout, outdoor seatings etc. and health and safety measures such as “masks required” and “social distancing enforced ”, etc. Based on the original fields obtained, I derived a few key variables which anchored my subsequent analysis:

  • Recent Rating (avg) - Average star ratings of restaurants since 3/1/2020, which is roughly when the Pandemic lockdown start affecting restaurants across the U.S.
  • Difference between Recent Rating and Overall Rating – To understand how ratings has changed before and after the pandemic started.
  • Weekly Total Business Hours – To understand if there is a relationship between length of business operation and recent review ratings.
  • Delivery, Takeout, Curbside Pickup, Outdoor Seating, Safety Measures – Each of those has been converted to a Boolean variable indicating whether the restaurant does or does not have the service available. Note that Safety Measures include any of the following: “Mask required”, “Staff wears masks”, “Staff wears gloves”, “Limited capacity”, “Hand sanitizer provided”, “Social distancing enforced”, “Contactless payment”, “Temperature checks”.

Data Analysis

All Restaurants – Recent Average Rating vs. Cumulative Average Rating

To start the analysis, I compared the overall cumulative star ratings with the recent average ratings. Again the recent rating is looking at reviews since March 2020 when the lockdown starts. At a glance we can see over half of the restaurants have an average 4 star review, with kind of an even spread of 3.5 and 4.5. The recent star ratings since March 1st is more skewed to the left, with more people giving higher reviews.

Comparing the two average ratings, the recent average ratings since Covid19 started (4.08) is actually 0.1 higher than the cumulative average rating (3.99), which is a little surprising to me because I was expecting more of a decrease with all the lockdown, store closing, and some restaurants who were not used to do delivery having to completely change the way they operate.

I do want to point out the limitation of using star ratings as the sole measure of business outcome, obviously this analysis would be more complete if I were able to get the restaurants’ sales volume or revenue and compare that through time, but due to time and resource limitation, I just used the star ratings as a proxy of restaurants’ performance, which is also a very important performance indicator. 

Data Analysis on Covid-19's Impact on the Restaurant Business

All Restaurants – Comaprison by Service Availability

Next, I examined the overall service availability - about 80% of the restaurants offer takeout, 75% offers delivery, 20% curbside pickup, and only 15% of the restaurants have safety measures, and only 10% outdoor seating and 10% sit-down dining available. Interestingly, I did not see much of a difference in recent ratings with delivery YN and takeout YN, however that is not to say that those are not important, since delivery and takeout probably account for the majority of all restaurants sales right now, and my assumption is that some restaurants just did not put those service labels in their Yelp homepage.

 On the other hand, restaurants with outdoor seating availability and safety measures are generally higher than those that don't have them, and the recent rating difference is about 0.1 as shown on those box plots respectively. Even though it doesn’t seem like that much of a difference from the graph, I conducted a two sample t-test to see whether the difference in average recent rating for restaurants that do or do not have safety measures and outdoor seating are statistically significant. The resulting p-values for both tests are very small, meaning that difference in rating are statistically significant. Data Analysis on Covid-19's Impact on the Restaurant Business

All Restaurants – Comparison by Price Tier

I then went on to explore restaurants across different price range. One interesting trend I noticed is that restaurants with the higher the price tag actually have received higher recent average ratings. I also compared the total weekly hours of operations for the different price tiers, and found that the more expensive restaurants generally operate in shorter hours, and have more outdoor seating options available. 

Data Analysis on Covid-19's Impact on the Restaurant Business

Comparison by Cuisine Types

Now let’s break down the data by different types of cuisines. Among the 7 categories, while all cuisines had an improvement in recent reviews, we can see that comparing the cumulative average and recent ratings, Chinese food had the most improvement recently, and Italian food also had a lot of improvement while also having the highest recent ratings. Mediterranean food, while having the highest cumulative average ratings, had the least improvement during the pandemic.  


Digging a little deeper into percentage of services available across cuisine types, it seems like Thai food offers the most delivery and takeout, and American restaurants offer the most outdoor seating, sit-down dining, curbside pickup, and are the best at implementing safety measures.

Safety Measures

Then I further looked into how did the two important factors: safety measure and outdoor seating options affected each cuisine types, and found out that while those measures benefited most cuisines, Mexican restaurants had the most noticeable increase in ratings by implementing those services.

Outdoor Seating

Comparison by Locations

I did a similar analysis to break down the data by three different locations  - NYC, San Francisco, Austin TX. The result showed that SF had the most improvement in recent ratings compared with the cumulative average, while Austin had the least improvement. NYC has the highest review ratings throughout. Service availability wise, NYC has the most delivery/takeout services, Austin has the most sit-down dining, and SF has the most safety measures in place.


Overall: The recent (3/1/2020 ~ now) Yelp rating among those restaurants still operating during the pandemic has generally been higher than the cumulative rating (+0.1). However, it definitely does not mean restaurants are doing better than before. Given the fact that thousands of restaurants having to drastically adjust their operations to accommodate to city guidelines, people more or less appreciate the services provided by restaurants during this special time and want to offer more support.

Among the 4800 restaurants representing 7 cuisines across 3 major U.S. cities, 80% of them now offer delivery and takeout services. Safety measures on the other hand are still falling short, with only 15% of the restaurants have any health/safety related information listed on their Yelp postings.

By Cuisine: Italian restaurants have the highest average rating during the pandemic, Chinese restaurants has had the most improvement

By City: NYC restaurants have remained the highest ratings throughout, while San Francisco has had the most improvement recently. San Francisco also has the most safety measures available comparing with the other two cities.

My recommendations for the restaurant owners would be:

  • Try to implement as many health and safety measures as possible, and make sure to list them on the Covid-19 Updates section on Yelp. Those measures will help boost customers trust when mediate their safety concerns when choosing a place to dine with.
  • If conditions permit, opening up some outdoor seats would mostly like help.

Future Work

  • Find data about restaurants’ volume / revenue at different times and combine it with review ratings to better indicate restaurants’ business overall outcome
  • Use linear regression / other ML models to uncover factors that correlates to recent review rating
  • Explore demographic characteristics in NYC, SF and Austin and cross-reference
  • Further explore patterns around restaurants sub-categories and find interesting patterns. (E.g. Ramen, fast-food, cafĂ©, tapas, etc.)


About Author

Qing Ying

Qing (Sophie) Ying graduated from UC Berkeley with a Master's degree in Industrial Engineering and Operations Research. She has been a product manager in a healthcare technology startup for 3 years, where she developed various data analytics products...
View all posts by Qing Ying >

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 2020 Revenue 3-points agriculture air quality airbnb airline alcohol Alex Baransky algorithm alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus ames dataset ames housing dataset apartment rent API Application artist aws bank loans beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep boston safety Bundles cake recipe California Cancer Research capstone car price Career Career Day citibike classic cars classpass clustering Coding Course Demo Course Report covid 19 credit credit card crime frequency crops D3.js data data analysis Data Analyst data analytics data for tripadvisor reviews data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization database Deep Learning Demo Day Discount disney dplyr drug data e-commerce economy employee employee burnout employer networking environment feature engineering Finance Financial Data Science fitness studio Flask flight delay gbm Get Hired ggplot2 googleVis H20 Hadoop hallmark holiday movie happiness healthcare frauds higgs boson Hiring hiring partner events Hiring Partners hotels housing housing data housing predictions housing price hy-vee Income Industry Experts Injuries Instructor Blog Instructor Interview insurance italki Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter las vegas airport lasso regression Lead Data Scienctist Lead Data Scientist leaflet league linear regression Logistic Regression machine learning Maps market matplotlib Medical Research Meet the team meetup methal health miami beach movie music Napoli NBA netflix Networking neural network Neural networks New Courses NHL nlp NYC NYC Data Science nyc data science academy NYC Open Data nyc property NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time performance phoenix pollutants Portfolio Development precision measurement prediction Prework Programming public safety PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R Data Analysis R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn seafood type Selenium sentiment analysis sentiment classification Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau teachers team team performance TensorFlow Testimonial tf-idf Top Data Science Bootcamp Top manufacturing companies Transfers tweets twitter videos visualization wallstreet wallstreetbets web scraping Weekend Course What to expect whiskey whiskeyadvocate wildfire word cloud word2vec XGBoost yelp youtube trending ZORI