Indepth of Arabican Coffee over the world

Background

Coffee is a brewed drink prepared from roasted coffee beans, the seeds of berries from certain Coffea species. When coffee berries turn from green to bright red in color – indicating ripeness – they are picked, processed, and dried.Dried coffee seeds (referred to as “beans”) are roasted to varying degrees, depending on the desired flavor. Roasted beans are ground and then brewed with near-boiling water to produce the beverage known as coffee.

Coffee is darkly colored, bitter, slightly acidic and has a stimulating effect in humans, primarily due to its caffeine content. It is one of the most popular drinks in the world, and can be prepared and presented in a variety of ways (e.g., espresso, French press, caffè latte). It is usually served hot, although iced coffee is common.

The two most commonly grown coffee bean types are C. arabica and C. robusta. Coffee plants are now cultivated in over 70 countries, primarily in the equatorial regions of the Americas, Southeast Asia, the Indian subcontinent, and Africa. As of 2018, Brazil was the leading grower of coffee beans, producing 35% of the world total. Coffee is a major export commodity as the leading legal agricultural export for numerous countries. It is one of the most valuable commodities exported by developing countries. Green, unroasted coffee is one of the most traded agricultural commodities in the world. The way developed countries trade coffee with developing nations has been criticised, as well as the impact on the environment with regards to the clearing of land for coffee-growing and water use. Consequently, the markets for fair trade and organic coffee are expanding.

At this time i will explore more about Arabica Coffee, a type of coffee that can produce a tasty coffee based on the soil around it. We will find out the cluster or so we called share the same similarities of the taste.

First, prepare the libraries

library(tidyverse)
library(tidyr)
library(dplyr)
library(GGally)
library(FactoMineR)
library(factoextra)
library(animation)
library(factoextra)

Data Preprocess

Selecting the important variable and mutate them into integer.

coffee <- read.csv("datasets_113109_269905_Coffee-modified.csv")
coffee1 <- coffee %>%
  select(Country.of.Origin, Aroma, Flavor, Aftertaste, Acidity, Body, Balance, Uniformity, Clean.Cup, Sweetness, Moisture, Quakers, Color) %>% 
  mutate(Aroma = as.integer(Aroma),
         Flavor = as.integer(Flavor),
         Aftertaste = as.integer(Aftertaste),
         Acidity = as.integer(Acidity),
         Body = as.integer(Body),
         Balance = as.integer(Balance),
         Uniformity = as.integer(Uniformity),
         Clean.Cup = as.integer(Clean.Cup),
         Sweetness = as.integer(Sweetness),
         Moisture = as.integer(Moisture),
         Quakers = as.integer(Quakers),
         Color = as.integer(Color)) %>%
          na.omit()
coffeeagg <- aggregate(x = list(Aroma = coffee1$Aroma,
                                Flavor = coffee1$Flavor,
                                Aftertaste = coffee1$Aftertaste,
                                Acidity = coffee1$Acidity,
                                Body = coffee1$Body,
                                Balance = coffee1$Balance,
                                Uniformity = coffee1$Uniformity,
                                Clean.Cup = coffee1$Clean.Cup,
                                Sweetness = coffee1$Sweetness,
                                Moisture = coffee1$Moisture,
                                Quakers = coffee1$Quakers,
                                Color = coffee1$Color),
                       by = list(Country.of.Origin = coffee1$Country.of.Origin), FUN = "mean")

coffeeagg1 <- coffeeagg %>%
  column_to_rownames(var = "Country.of.Origin")
head(coffeeagg1)

Now we have a clean data and ready to be explored.

Finding optimum K

fviz_nbclust(coffeeagg1, kmeans, method = "wss")

From the graph above, its slightly lean on number 6 so i will use it as clustering optimal number.

Clustering

RNGkind(sample.kind = "Rounding")

## Warning in RNGkind(sample.kind = "Rounding"): non-uniform 'Rounding' sampler
## used

set.seed(500)
coffee_cluster <- kmeans(coffeeagg1, 8)

Visualize

fviz_cluster(object = coffee_cluster, data = coffeeagg1)

These are the cluster of each Arabica Coffee from around the world.

Lets check what other country that share the similarities of Indonesian’s beans

coffeeagg1 %>% 
  rownames_to_column("Country.of.Origin") %>% 
  mutate(cluster = as.factor(coffee_cluster$cluster)) %>% 
  filter(Country.of.Origin == "Indonesia") %>% 
  select(cluster)

coffeeagg1 %>% 
  rownames_to_column("Country.of.Origin") %>% 
  mutate(cluster = as.factor(coffee_cluster$cluster)) %>% 
  filter(cluster == 1) %>% 
  select(Country.of.Origin)

As you can see, our country, Indonesia, share the similarities with top beans producer around the globe, let say Guatemala, Tanzania and Brazilian coffee. (We know that our coffee has a good popularity around the globe (We dont talk about Luwak Coffee here, i think it’s not a natural process anymore :) ), even in Starbucks, the biggest Coffee-Chain in the world has release Toraja Sapan and Sumatra single origin for a long time, it means that our beans has big fans in this coffee industries.

Lets check the characteristic of each coffee cluster by cluster profiling

coffee_profiling <- coffeeagg1 %>% 
  mutate(cluster = as.factor(coffee_cluster$cluster)) %>% 
  group_by(cluster) %>% 
  summarise_all(mean)

Profiling

coffee_profiling %>% 
  pivot_longer(cols = -1, names_to = "type", values_to = "score") %>% 
  ggplot(aes(x = cluster, y = score))+
  geom_col(aes(fill = cluster))+
  facet_wrap(~type)

Looks like Cluster 8 has the highest amount of Acidity, Aftertaste, Aroma and Flavor. Lets observe from what country they are!

coffeeagg1 %>% 
  rownames_to_column("Country.of.Origin") %>% 
  mutate(cluster = as.factor(coffee_cluster$cluster)) %>% 
  filter(cluster == 8) %>% 
  select(Country.of.Origin)

Well, looks like 3 countries are the top in the market, i believe Ethiopia is great in producing coffee beans, i have been tasting Ethiopia’s Sidamo Guji and Yirgacheffe and they tasted awesome, it is light as some Arabican in common but you can’t find the Boldness there, but i am not quite sure that United States and PNG produce a good descent coffee beans as many of us know that perhaps they just do a bean trading there :).

Lets check for the most expensive coffee beans in the world, Panama Geisha.

Why it become the most expensive beans in the world? The difficulties to grow the coffee plant by its geographic location and the soil also the challenging climate all over the year led the Speciality Coffee Association (SCA) gives it the highest score and all the coffee drinker seems like to look after it anywhere, that’s why the price are getting high. It’s quite rare to find the beans here, i have been tasting one from Ombe Kofie, Jakarta.

Lets check the cluster of Panama Geisha

coffeeagg1 %>% 
  rownames_to_column("Country.of.Origin") %>% 
  mutate(cluster = as.factor(coffee_cluster$cluster)) %>% 
  filter(Country.of.Origin == "Panama") %>% 
  select(cluster)

Surprisingly it is lower by almost overal mark from cluster 7, perhaps they don’t get the best beans during this time :)

My favourite beans from Indonesia are Pangalengan, Malabar, Gambung, Toraja, Kalosi Aceh and Kintamani, they are all taste completely different because of soil and plantation that been planted around or before it.

Conclusion

Like i said, Arabican beans are usually light in Body but strong in acidity and flavor, some says that Arabican are better than Robusta, vice versa, but it depends on you, there’s nothing wrong with it. You can just tailor it personally with your own method, V60, Chemex, Aeropress and Espresso are all good as long as you know what taste they will produce after the extraction process.

I hope many of us will appreciate all of the coffee varieties from around the world and can exactly know how to grasp the best potential of each, from the plantation to extraction method, and now please enjoy your coffee!