Introduction

Hello everyone! All people want to eat at good and high rated resataurants. Therefore, we all have to know which restaurant is good,better, best or even worst. Regarding to this situation, an institution called Michelin developed a way to score and evaluate how good a restaurant is. They give stars to a restaurant which is good for them. More stars that a restaurant get more good it will be.

This document, summarise the distribution of 1,2, and 3 stars restaurant in many different countries. However, after observing the data, we can conclude that there are only 3 different continents who have 1,2,3 star restaurants.

At the end, I hope that you can get much information abaout the distiburion of michelin restaurant in the world and someday it can be a reference to go. Thankyou and happy reading!

Preparation

Before start, we need to load our packages

library(ggplot2)
library(plotly)
library(dplyr)
library(glue)
library(plyr)
library(tidyr)

Data Processing

1-Star

onestar <- read.csv(file = "one-star-michelin-restaurants.csv")[ , -c(3,4,5,7,10)]

Check Data Types

I will check for the data types

str(onestar) #check for all data types

## 'data.frame':    549 obs. of  5 variables:
##  $ name   : Factor w/ 544 levels "108","21212",..: 228 356 135 77 124 112 367 521 411 497 ...
##  $ year   : int  2019 2019 2019 2019 2019 2019 2019 2019 2019 2019 ...
##  $ region : Factor w/ 24 levels "Austria","California",..: 1 1 1 1 1 1 1 1 1 1 ...
##  $ cuisine: Factor w/ 67 levels "American","Asian",..: 16 12 16 38 45 45 16 12 35 67 ...
##  $ price  : Factor w/ 6 levels "$","$$","$$$",..: 5 5 5 5 4 5 5 5 5 5 ...

Convert as.factor

I found that “year” coloumn still in int, therefore I am going to convert it to factor

onestar$year <- as.factor(onestar$year)

Got it! Now I will check for the summary of the region, so I can group the region into five different continents in the world

summary(onestar[ , "region"])

##        Austria     California        Chicago        Croatia Czech Republic 
##             12             69             18              5              2 
##        Denmark        Finland         Greece      Hong Kong        Hungary 
##             22              6              3             44              5 
##        Ireland          Macau  New York City         Norway         Poland 
##             13             11             55              7              2 
## Rio de Janeiro      Sao Paulo      Singapore    South Korea         Sweden 
##              5             10             34             19             16 
##         Taipei       Thailand United Kingdom  Washington DC 
##             18             22            138             13

Making new coloumn

Making a new column named “continent”

asia <- c("Hong Kong", "Macau","Singapore","South Korea","Taipei","Thailand")
america <- c("California","Chicago","New York City","Washington DC", "Rio de Janeiro","Sao Paulo")

onestar$continent <- ifelse(onestar$region %in% asia , "Asia", ifelse(onestar$region %in% america, "America", "Europe"))
  
onestar$continent <-as.factor(onestar$continent)

Using ggplot2

Let us see the diagram of all the number of one-michelin star restaurant in the world

ggplot(data = onestar, mapping = aes(x = continent)) +
  geom_bar(mapping = aes(fill = continent, 
                         )) +
  labs (title = "Numbers of One Michelin Star Restaurants", x="", y="Counts", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white"))

Wohoo! Most of one-michelin star restaurants are located in Europe!

Now let’s compare the prices between the continents

table (onestar$price, onestar$continent)

##        
##         America Asia Europe
##   $           0   29      0
##   $$         19   40      8
##   $$$        50   49     30
##   $$$$       87   19     18
##   $$$$$      14   11     24
##   N/A         0    0    151

onestar$rate <- ifelse(onestar$price == "$", 1 , 
                       ifelse(onestar$price == "$$", 2 ,
                              ifelse(onestar$price == "$$$", 3,
                                     ifelse(onestar$price== "$$$$", 4 , 5))))
table ( onestar$continent, onestar$rate)

##          
##             1   2   3   4   5
##   America   0  19  50  87  14
##   Asia     29  40  49  19  11
##   Europe    0   8  30  18 175

table_rate <-aggregate(rate ~ continent, onestar, FUN= mean)
table_rate$rate <- round(table_rate$rate, 2)



p1<- ggplot(data = table_rate, mapping = aes(x = continent, y= rate)) +
  geom_col(mapping = aes(fill = continent, 
  text = glue("Continent : {continent} 
              Rate : {rate}"), 
           )) +
  labs (title = "Rate Comparison Between One Michelin Star Restaurant", x="", y="Rate", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white")) 
    
ggplotly(p1, tooltip = "text")

One star restaurants in Europe are the most expensive than the others.

Now let’s look at 2 stars restaurants

2-Star

twostar<- read.csv(file="two-stars-michelin-restaurants.csv")[ , -c(3,4,5,7,10)]
twostar$year <- as.factor(twostar$year)
str(twostar)

## 'data.frame':    110 obs. of  5 variables:
##  $ name   : Factor w/ 108 levels "Ã‰criture","aâ\200šoâ\200šc",..: 80 37 62 46 84 89 12 19 15 50 ...
##  $ year   : Factor w/ 2 levels "2018","2019": 2 2 2 2 2 2 2 2 2 2 ...
##  $ region : Factor w/ 19 levels "Austria","California",..: 1 1 1 1 1 1 2 2 2 2 ...
##  $ cuisine: Factor w/ 28 levels "Californian",..: 7 7 7 22 22 7 5 5 20 5 ...
##  $ price  : Factor w/ 6 levels "$","$$","$$$",..: 5 5 5 5 5 5 4 4 4 4 ...

Making continent coloumn

We do the same thing as one star

twostar$continent <- ifelse(twostar$region %in% asia , "Asia", ifelse(twostar$region %in% america, "America", "Europe"))
  
twostar$continent <-as.factor(twostar$continent)

Creating ggplot2

And How about 2 star?

ggplot(data = twostar, mapping = aes(x = continent)) +
  geom_bar(mapping = aes(fill = continent)) +
  labs (title = "Numbers of Two Michelin Star Restaurants", x="", y="Counts", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white"))

Just like one star, mots of two stars michelin restaurants are located in Europe! (38)

table (twostar$price, twostar$continent)

##        
##         America Asia Europe
##   $           0    2      0
##   $$          0    6      0
##   $$$         0   11      1
##   $$$$       33    8      9
##   $$$$$       3    9      8
##   N/A         0    0     20

twostar$rate <- ifelse(twostar$price == "$", 1 , 
                       ifelse(twostar$price == "$$", 2 ,
                              ifelse(twostar$price == "$$$", 3,
                                     ifelse(twostar$price== "$$$$", 4 , 5))))

table ( twostar$continent, twostar$rate)

##          
##            1  2  3  4  5
##   America  0  0  0 33  3
##   Asia     2  6 11  8  9
##   Europe   0  0  1  9 28

table_rate_two <-aggregate(rate ~ continent, twostar, FUN= mean)
table_rate_two$rate <- round(table_rate_two$rate, 2)



p2<- ggplot(data = table_rate_two, mapping = aes(x = continent, y= rate)) +
  geom_col(mapping = aes(fill = continent, 
  text = glue("Continent : {continent} 
               Rate : {rate}") )) +
  labs (title = "Rate Comparison Between Two Michelin Star Restaurant", x="", y="Rate", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white")) 

ggplotly(p2, tooltip = "text")

And again, two star restaurants in Europe are the most expensive, while the cheapest are in asia.

3-Star

threestar<- read.csv(file = "three-stars-michelin-restaurants.csv")[ , -c(3,4,5,7,10)]
names(threestar)

## [1] "name"    "year"    "region"  "cuisine" "price"

threestar$year <- as.factor(threestar$year)
str(threestar)

## 'data.frame':    36 obs. of  5 variables:
##  $ name   : Factor w/ 36 levels "8Â½ Otto e Mezzo - Bombana",..: 4 23 6 26 5 33 35 28 3 14 ...
##  $ year   : Factor w/ 1 level "2019": 1 1 1 1 1 1 1 1 1 1 ...
##  $ region : Factor w/ 13 levels "Austria","California",..: 1 2 2 2 2 2 2 2 3 4 ...
##  $ cuisine: Factor w/ 16 levels "American","Asian",..: 7 6 2 6 6 6 6 6 6 7 ...
##  $ price  : Factor w/ 5 levels "$$","$$$","$$$$",..: 4 3 3 3 3 3 3 3 3 3 ...

Creating ggplot2

And last for the 3 star restaurants, are there any differences?

threestar$continent <- ifelse(threestar$region %in% asia , "Asia", ifelse(threestar$region %in% america, "America", "Europe"))
  
threestar$continent <-as.factor(threestar$continent)

ggplot(data = threestar, mapping = aes(x = continent)) +
  geom_bar(mapping = aes(fill = continent)) +
  labs (title = "Numbers of Three Michelin Star Restaurants", x="", y="Counts", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white"))

Wow! Although Europe has the highest numbers of one-star and two-star restaurants in the world, it has the lowest in three star restaurants. The highest is now in America while Asia stands in 2nd. Interesting!

Now how about the prices?

table (threestar$price, threestar$continent)

##        
##         America Asia Europe
##   $$          0    2      0
##   $$$         0    2      0
##   $$$$       14    6      3
##   $$$$$       0    3      1
##   N/A         0    0      5

threestar$rate <- ifelse(threestar$price == "$", 1 , 
                       ifelse(threestar$price == "$$", 2 ,
                              ifelse(threestar$price == "$$$", 3,
                                     ifelse(threestar$price== "$$$$", 4 , 5))))

table ( threestar$continent, threestar$rate)

##          
##            2  3  4  5
##   America  0  0 14  0
##   Asia     2  2  6  3
##   Europe   0  0  3  6

table_rate_three <-aggregate(rate ~ continent, threestar, FUN= mean)
table_rate_three$rate <- round(table_rate_three$rate, 2)



p3<- ggplot(data = table_rate_three, mapping = aes(x = continent, y= rate)) +
  geom_col(mapping = aes(fill = continent, 
  text = glue("Continent : {continent} 
               Rate : {rate}"))) +
  labs (title = "Rate Comparison Between Three Michelin Star Restaurant", x="", y="Rate", fill = "Continent") +
  theme (plot.title = element_text(face = "bold"), panel.grid = element_blank(), panel.background = element_rect(fill = "white"))


ggplotly (p3, tooltip= "text")

As wee see, from one,two,and three star restaurants, Europe has the most expensives dishes in the world.

Creating chart to visualize all data

Now, for a better visualization we make the charts simpler.

asia1 <- nrow(onestar[onestar$continent == "Asia", ])
asia2 <- nrow(twostar[twostar$continent == "Asia", ])
asia3 <- nrow(threestar[threestar$continent == "Asia", ])

asia.mul <- rbind (asia1, asia2, asia3)
asia.mul <- as.data.frame(asia.mul)
names(asia.mul)[names(asia.mul) == "V1"] <- "Asia"
asia.mul$Star <- c("1-Star", "2-Star","3-Star")
asia.mul<- asia.mul[, c(2,1)]
asia.mul$Star <- as.factor(asia.mul$Star)
asia.mul

##         Star Asia
## asia1 1-Star  148
## asia2 2-Star   36
## asia3 3-Star   13

america1 <- nrow(onestar[onestar$continent == "America", ])
america2 <- nrow(twostar[twostar$continent == "America", ])
america3 <- nrow(threestar[threestar$continent == "America", ])
america.mul <- rbind (america1, america2, america3)
america.mul <- as.data.frame(america.mul)
names(america.mul)[names(america.mul) == "V1"] <- "America"
america.mul$Star <- c("1-Star", "2-Star","3-Star")
america.mul<- america.mul[, c(2,1)]
america.mul$Star <- as.factor(america.mul$Star)
america.mul

##            Star America
## america1 1-Star     170
## america2 2-Star      36
## america3 3-Star      14

europe1 <- nrow(onestar[onestar$continent == "Europe", ])
europe2 <- nrow(twostar[twostar$continent == "Europe", ])
europe3 <- nrow(threestar[threestar$continent == "Europe", ])
europe.mul <- rbind (europe1, europe2, europe3)
europe.mul <- as.data.frame(europe.mul)
names(europe.mul)[names(europe.mul) == "V1"] <- "Europe"
europe.mul$Star <- c("1-Star", "2-Star","3-Star")
europe.mul<- europe.mul[, c(2,1)]
europe.mul$Star <- as.factor(europe.mul$Star)
europe.mul

##           Star Europe
## europe1 1-Star    231
## europe2 2-Star     38
## europe3 3-Star      9

mat<- cbind (asia.mul,america.mul,europe.mul)
mat<- mat [, -c(3,5)]
mat

##         Star Asia America Europe
## asia1 1-Star  148     170    231
## asia2 2-Star   36      36     38
## asia3 3-Star   13      14      9

mulp<- pivot_longer(mat, cols= c("Asia","America","Europe"))
mulp

## # A tibble: 9 x 3
##   Star   name    value
##   <fct>  <chr>   <int>
## 1 1-Star Asia      148
## 2 1-Star America   170
## 3 1-Star Europe    231
## 4 2-Star Asia       36
## 5 2-Star America    36
## 6 2-Star Europe     38
## 7 3-Star Asia       13
## 8 3-Star America    14
## 9 3-Star Europe      9

colnames(mulp) <- c("Star", "Continent","Star_Count")
mulp$Continent <- as.factor (mulp$Continent)
mulp

## # A tibble: 9 x 3
##   Star   Continent Star_Count
##   <fct>  <fct>          <int>
## 1 1-Star Asia             148
## 2 1-Star America          170
## 3 1-Star Europe           231
## 4 2-Star Asia              36
## 5 2-Star America           36
## 6 2-Star Europe            38
## 7 3-Star Asia              13
## 8 3-Star America           14
## 9 3-Star Europe             9

Time to ggplot!

library(extrafont)
p4 <- ggplot (data = mulp, mapping = aes(x= Continent, y=Star_Count)) +
  geom_col (mapping = aes(fill = Star, 
  text = glue("Continent : {Continent} 
               Star Count : {Star_Count}")), position = "dodge") +
  labs(x ="", y="Count", fill = "", title = "Number of Michelin Star Restaurant in America, Asia, and Europe") +
  theme (plot.title = element_text(size = 10, face = "bold"),  panel.background = element_rect(colour = "red", fill="beige"),axis.title.y = element_text(size = 10))
  
ggplotly(p4, tooltip = "text")

Chart above is the summary of total number of 1-star,2-star,3-star michelin restaurants in America, Asia , and Europe.

Trend in Michelin Restaurants

Our data have information about the time (in year) of michelin restaurants in the world. We will see if there are changes about the number of michelin restaurant year by year (starting from 2018)

Data Wrangling

We group the data by each continents.

#For Asia
oneasia<- onestar %>% 
  filter (continent =="Asia") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("1-star"))

twoasia<- twostar %>% 
  filter(continent == "Asia") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("2-star"))

threeasia<- threestar %>% 
  filter(continent == "Asia") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("3-star"))

#For America 
oneamerica<- onestar %>% 
  filter (continent =="America") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = "1-star") 



twoamerica<- twostar %>% 
  filter(continent == "America") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("2-star"))


threeamerica<- threestar %>% 
  filter(continent == "America") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("3-star"))

#For Europe

oneeurope<- onestar %>% 
  filter (continent == "Europe") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("1-star"))

twoeurope<- twostar %>% 
  filter(continent == "Europe") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("2-star"))

threeeurope<- threestar %>% 
  filter(continent == "Europe") %>% 
  select(year, continent) %>% 
  group_by(year, continent) %>%
  count() %>% 
  mutate(star = as.factor("3-star"))

Data Visualisation

Table joining using full join

star1<- oneasia %>% 
  full_join(oneamerica) %>% 
  full_join(oneeurope)

star2<- twoasia %>% 
  full_join(twoamerica) %>% 
  full_join(twoeurope)

star3<- threeasia %>% 
  full_join(threeamerica) %>% 
  full_join(threeeurope)

fix<- star1 %>% 
  full_join(star2) %>% 
  full_join(star3) 

fix[,c("year", "star")] <- lapply(fix[,c("year", "star")], as.factor)

fix %>% 
  arrange (continent)

##    year continent freq   star
## 1  2019   America  170 1-star
## 2  2019   America   36 2-star
## 3  2019   America   14 3-star
## 4  2018      Asia   34 1-star
## 5  2019      Asia  114 1-star
## 6  2018      Asia    5 2-star
## 7  2019      Asia   31 2-star
## 8  2019      Asia   13 3-star
## 9  2019    Europe  231 1-star
## 10 2019    Europe   38 2-star
## 11 2019    Europe    9 3-star

Now we make the plot!

p10<- fix %>% 
  ggplot(mapping = aes(x= continent, y=freq))+
  geom_col(aes(fill = year, text = glue("Continent = {continent}
                                         Year : {year}
                                         Counts : {freq}")), position = "dodge")+
  facet_wrap(~star, nrow =1 ) +
  theme_minimal()+
  labs(title = "Trend of Michelin Star Restaurants Between 2018 and 2019", y = "counts")+
  theme(plot.title = element_text(size = 10, face = "bold"),
        axis.title.y = element_text(size = 11))

ggplotly(p10, tooltip ="text")

As we can see from the charts above, only Asia has michelin restaurants in 2018.

Conclusion

After processing all data, here what we can get:

1.Europe has the highest number of Michelin Stars restaurant. This indicates that most of European restaurants are highly rated.
2 The most expensive restaurants are located in Europe
3 No 1,2,3 stars Michelin restaurants established in Africa and Australia
4.Only Asia has michelin restaurants in 2018

So this this the end of the article.

Thank you for reading and see you on the next ones!

Michelin Restaurants Around The World

Alfado Dhusi Sembiring

17/1/2020

Introduction

Preparation

Data Processing

1-Star

Check Data Types

Convert as.factor

Making new coloumn

Using ggplot2

2-Star

Making continent coloumn

Creating ggplot2

3-Star

Creating ggplot2

Creating chart to visualize all data

Trend in Michelin Restaurants

Data Wrangling

Data Visualisation

Conclusion