The Michelin curse that even black-and-white chefs can't escape: the story analyzed in R

If the Black and White Chef: Culinary Class Warfarein the summer? The contestants are split into black and white for a tense culinary showdown, with Michelin-starred chefs among them.

But did you know there's a dark shadow behind that shiny star? Michelin stars are a dream come true for chefs, but sometimes they come with a heavy burden. That's why they're called 'The curse of the Michelin star', or something like that.

Today we'll explore the Michelin curse using R. Shocking study finds 40% of Michelin-starred New York restaurants will be closed by 2019is present. Black and White ChefLet's look at the data to uncover the secrets of why this is the outcome of their dreams. We'll analyze the survival rate of Michelin-starred restaurants visually, and walk through the R code line by line.

What is the Michelin Curse?

A Michelin star is more than just an accolade. Chefs put in countless hours of hard work to earn this one star, but the pressure doesn't stop once the honor is earned. From the moment a Michelin star is awarded, chefs and restaurants are held to a higher standard, and there's more money and pressure to meet those expectations.

After being awarded a Michelin star, the cost of ingredients rises, chefs demand higher wages, and even customers' expectations seem to rise endlessly. Restaurants that don't survive this are often forced out of business, which is why people call the star both an honor and a curse.

Visualize the Michelin Curse with R

Now, let's analyze the data to see how much of an impact the Michelin curse really had. We'll use R to visualize the survival rate of Michelin-starred restaurants. I'll explain the code used in this process in detail, line by line, so even beginners can easily follow along!

# Load the required library
library(survival)
library(survminer)

# Create data of Michelin-starred restaurants from 2000 to 2014
# close_year: the year the restaurant closed, or NA if it is still open until 2019
data <- data.frame(
  year = c(2000:2014),
  close_year = c(2007, NA, 2012, NA, 2013, 2014, NA, 2017, 2018, 2019, NA, NA, 2019, NA, NA, NA)
)

# Calculate how many years a restaurant has been open
# If the restaurant is open until 2019, calculate how long it has been in operation as of 2019
data$time_open <- ifelse(is.na(data$close_year), 2019 - data$year, data$close_year - data$year)

# Indicate whether the restaurant is closed or not as a 1 or 0 (1: closed, 0: open)
data$status <- ifelse(is.na(data$close_year), 0, 1)

In fact, I tried to find the raw data from the Economist article as best I could, but I couldn't find it, so I can only visualize it at this level.

Code description:

1. data.frame: First, generate data on restaurants that received Michelin stars between 2000 and 2014. close_yearindicates the year the restaurant closed, and those still in business through 2019 are marked with a NAto handle this.
2. time_open: Code to calculate how many years the restaurant has been in business. If the restaurant is open until 2019, then we calculate the number of years of operation based on 2019.
3. status: A variable indicating whether the business is closed. It will be 1 if the business is closed and 0 if it is still in business.

If you check the data we created above, you'll see that it looks like this

Generate a survival curve by fitting a # Kaplan-Meier survival model
fit <- survfit(Surv(time_open, status) ~ 1, data = data)

Visualize the # survival curve (passing the data explicitly)
ggsurvplot(fit,
           data = data, # Pass data to ggsurvplot
           conf.int = TRUE, # Show confidence intervals
           xlab = "Years since receiving a Michelin star", # x-axis labels
           ylab = "Percentage of restaurants that are open", # y-axis label
           title = "Survival curve of Michelin-starred restaurants", # Graph title
           palette = "blue", # Graph color
           risk.table = TRUE, # Add risk table for each year
           ggtheme = theme_minimal()) # Apply minimal theme

Code description:

1. Kaplan-Meier modelto draw a survival curve for a restaurant. This survival curve visualizes how many years a restaurant has survived since receiving a Michelin star.
2. ggsurvplotA function to visualize the survival curve. conf.int = TRUEshows the confidence interval, risk.tableto further show the number of restaurants remaining in each year.
3. palette = "blue": I used blue to make the graph look better.

Explain the visualization results:

1. Survival curve overview:

• The blue line shows the survival curve of Michelin-starred restaurants from 2000 to 2014.
• The x-axis shows the number of years since the restaurant was awarded a Michelin star.
• The y-axis represents the percentage of restaurants that are still open.

2. Key points on the curve:

• The curve starts at 100% (or 1.0) at year 0, which means that all restaurants were in business immediately after receiving a Michelin star.
• Over time, the curve cascades down, indicating that some restaurants have gone out of business at that point in time.
• At points like 5 years, 7 years, and so on, the curve drops sharply, meaning that many restaurants have closed at that point.
• At the end of the graph (about 17 years), only about 251 TP3T of restaurants are still in business.

3. Shades of gray (confidence intervals):

• The gray shading around the blue line represents the confidence interval.
• This represents uncertainty about the probability of survival, with wider shading indicating greater uncertainty about that probability.
• Confidence intervals widen over time, which means that the uncertainty in the data increases over time.

4. Number at Risk table:

• At the bottom of the graph is a table called "Number at risk".
• This shows how many restaurants are still open at each point in time.
• For example, at year 0, there were 15 restaurants, and at year 10, there were only 7 restaurants left.
• After 17.5 years, only 1 restaurant remains in business.

5. Curve analysis:

• The sharp dip in the curve indicates when a large number of restaurants went out of business.
• After 10 years, the curve slowly descends, showing that fewer restaurants are going out of business at that point, but that closures are still happening.
• The final results show that only about 251 TP3T of restaurants are still in business after about 17.5 years.

The Michelin Curse, a heavy reality

The results of Michelin stars, as seen through survival curves, are clear. We can see that the survival rate of Michelin-starred restaurants drops sharply with each passing year, with many restaurants closing. In particular, we can see that the In New York, a shocking 40% of Michelin-starred restaurants will be closed by 2019is present.

Why are so many restaurants going out of business? It's because they're not using Higher customer expectations and rising costs That's because customers expect the best service and the best ingredients from a Michelin-starred establishment, and restaurants have to constantly invest to meet those expectations. But they can't keep up with the financial strain, and eventually go out of business.

Earning a Michelin star is undoubtedly the highest honor a chef can receive, but it also comes with a heavy responsibility. That's why we call it the Michelin curse. Not many chefs can escape this curse. Black and White Chefis one such case in point.

Recap: The Michelin curse that even black-and-white chefs can't escape

Today we're analyzing the Michelin curse with R, and the results are surprising. A Michelin star is an honor, but it comes with a lot of pressure and high expectations, and many restaurants close their doors because of it. Our data showed that this curse is not a myth, but a reality.

If you're interested in learning more about the Michelin Curse, you can run the R code we've presented today. And if the Kaplan-Meier model isn't something you're familiar with, I'll be back with another post to give you an easy introduction to it.

#Full R-Code

We'll end this post by showing the full code again.

# Load the required library
library(survival)
library(survminer)

# Create data of Michelin-starred restaurants from 2000 to 2014
# close_year: the year the restaurant closed, NA if it is still open until 2019
data <- data.frame(
    year = c(2000:2014),
    close_year = c(2007, NA, 2012, NA, 2013, 2014, NA, 2017, 2018, 2019, NA, NA, 2019, NA, NA, NA)
)

# Calculate how many years a restaurant has been in operation
# If the restaurant is open until 2019, calculate the number of years of operation based on 2019
data$time_open <- ifelse(is.na(data$close_year), 2019 - data$year, data$close_year - data$year)

# Indicate whether the restaurant is closed or not as a 1 or 0 (1: closed, 0: open)
data$status <- ifelse(is.na(data$close_year), 0, 1)

# Generate a survival curve by applying the Kaplan-Meier survival model
fit <- survfit(Surv(time_open, status) ~ 1, data = data)

Visualize the # survival curve (passing the data explicitly)
ggsurvplot(fit,
           data = data, # Pass data to ggsurvplot
           conf.int = TRUE, # Show confidence intervals
           xlab = "Years since receiving a Michelin star", # x-axis labels
           ylab = "Percentage of restaurants that are open", # y-axis label
           title = "Survival curve of Michelin-starred restaurants", # Graph title
           palette = "blue", # Graph color
           risk.table = TRUE, # Add risk table for each year
           ggtheme = theme_minimal()) # Apply minimal theme