Beer, standing as the world’s most widely consumed alcoholic drink, takes the third spot in overall beverage consumption, surpassed only by water and tea 1. However, as a beverage enjoyed worldwide, we believe that beer isn’t just a drink - but a cultural guide. People everywhere sip on this frothy concoction, and the way they enjoy it can vary from place to place and change as the years go by. In this study, we aim to find out exactly what these different preferences are and their evolutions! More specifically, we will identify the beer preferences for each location and analyze how their evaluation and spreading behave over time. Parallel to that, we will zoom out a bit and investigate the “global” beer preferences, which will give us further insights into what the evolution of beer preferences signifies on a larger scale. Furthermore, we want to see which locations exhibit “similar” beer preferences, and how these similarities change over time. In the end, we will try to understand how these similarities and differences are influenced by two factors: the geographical distances and the differences in wealth.

The dataset

The data used for this study is collected from two different beer rating websites, namely BeerAdvocate and RateBeer. The dataset coming from each website includes information about different beers, their corresponding breweries, users, and the ratings that users gave to the beers. We first combined the two datasets into a single dataset and then filtered out the years and locations that offered too small of data to allow for reliable conclusions. In the end, our study is concerned only with the below-given time range and locations:

So, what does a ‘beer preference’ even mean?

Now the question arises: “What do we mean by a ‘beer preference’?”. We define the beer preference of a location based on two different criteria:

The first criterion is the preferred beer style and the second is the preferred location of the brewery. For example, if the ratings coming from the users in Luxembourg indicate that they prefer “Imperial Stout” as the beer style and they prefer to drink beers from breweries that are located in the Netherlands, then their beer preference is “Imperial Stout” in terms of the style and “Netherlands” in terms of the brewery location.

For each criterion (beer style or brewery location), we define the “preference” again based on two different criteria: popularity-wise or rating-wise. As the names indicate, for each location the popularity-based approach captures the beer style or brewery location that got the largest amount of ratings, whereas the rating-based approach captures the beer style or brewery location that got the highest average rating.

Since both definitions of preference have two different categories, we end up with four different metrics of beer preference. But don’t worry! To not overwhelm you, we collect each of our findings in groups of four and you will be able to access them using a simple dropdown menu!

Beer Preferences versus Location and Time

Now, let’s find out what the beer preferences of each location are and how they evolve over time. For this, we first group the beer styles and brewery locations under broader categories so that we can assign them to a meaningful set of colors. Then, we color each location on the world map based on its beer preference. You can use the dropdown menu to look at different metrics, the slider to navigate through the years, and hover your mouse over a location in the map to see its name!

Let’s evaluate our results for each metric briefly:

Ratings of beer styles:

We kick off with Belgian Ale reigning supreme in 2006. By 2009, Lambic & Sour Ale claims dominance, starting in North America and spreading to some European countries. Fast forward to 2011, and Lambic & Sour Ale takes control across nearly all locations in North America but loses ground in Europe. Up until 2014, the global beer map is a battleground between Lambic Sour Ale and Belgian Ale. Post-2015, Lambic & Sour Ale blankets the majority of the world. Notably, by 2017, only three styles—Lambic & Sour Ale, Belgian Ale, and Stout—shared the spotlight across the entire globe. It is also interesting to observe that Canada, Norway, and Sweden exhibit similar patterns in their preference shifts for the most part.

Popularity of beer styles:

We begin with Lager as the predominant beer worldwide, facing competition, particularly from Pale Ale in the U.S. and Belgian Ale in Europe. Up until 2011, the scenario remains relatively consistent, albeit with Pale Ale and Belgian Ale gaining ground in the U.S. and Europe, respectively. In 2012, Lager significantly diminished in popularity in the U.S., and Europe became divided among Pale Ale, Pilsener, and Belgian Ale, with Pale Ale emerging as the globally most popular. While the situation remains relatively stable until 2017, there’s a notable rise in popularity for Lambic & Sour Ale in North America towards the end. Interestingly, it’s observed that some Nordic countries, including Iceland, Sweden, and Finland, closely align their preferences with those of North America.

Ratings of brewery locations:

In 2006, there existed only two preferred brewery locations in the world, namely Central/Western Europe and North America. Until 2010, breweries in North America gained some ground in Europe while breweries in Northern Europe started leading the rating race in North America and Australia instead of breweries in Central/Western Europe. From 2010 to 2017, we see that breweries in North America take over the world in terms of rating year by year.

Popularity of brewery locations:

We begin in 2006 with most of the locations consuming from their local breweries the most. However, it can be seen that some states of the U.S. and most of the countries of Europe prefer breweries from Central/Western Europe. As the years go by, we see that almost all locations become more and more “nationalist”, as almost every location consumes from their own local breweries. One interesting example of that are the European countries, mostly favoring breweries in Central/Western Europe in 2006, now drink mostly from the breweries from their specific location in Europe - East, North or South.

Global Beer Preferences

Okay cool, we can check out what kinds of beer different places like on the world map. But wouldn’t it also be interesting to see the “overall” opinion of the world and see how it changes over the years? However, we have to do this wisely. Simply looking at the total number of ratings and their distribution in that year would not represent the preference of the whole world since the locations with high populations heavily dominate the aggregated results.

Instead, we came up with a much more thoughtful and interesting approach. If the Eurovision Song Contest can do it, so can we! In the Eurovision Song Contest, each country gives points to their top 10 favorite songs. The song that gets the highest number of total points is the winner of the year.

We decided to do something similar. For each location, we assign points to their top 10 favorite beer styles or brewery locations. The beer style or brewery location that gets the highest number of points wins the contest. The points are assigned as follows: 12 points for the most preferred beer style or brewery location, 10 points for the second most preferred one, 8 points for the third most preferred one, and so on, the exact rating system used in the contest. This way, we obtain a better representation of the beer preferences of the whole globe. Note that we normally treat each U.S. state as a separate location. However, for the global beer preferences, we treat the whole U.S. as a single location or otherwise the U.S. would have too many votes in the contest.

Below, you can see the global trends versus time in the form of bar chart races! Select your metric of interest, pick your preference, and hope that it wins the contest!

Let’s again briefly evaluate our results for each metric:

Ratings of beer styles:

In general, Imperial Stout, Abt/Quadrupel, and Imperial IPA lead the race in the top 3 spots. However, it must be acknowledged that Lambic Style beers get rated higher and higher as the years go by, with four Lambic Style beers in the top 10 and Lambic Style - Gueuze in the second spot in 2017.

Popularity of beer styles:

We begin with Pilsener, Pale Lager, and Belgian Strong Ale taking up the first three spots in 2006. Pilsener lost a lot of popularity until 2012 when India Pale Ale, Imperial Stout, and American Pale Ale climbed the ranking ladder. Until 2017, India Pale Ale was the leader of the race, followed by American Pale Ale and Imperial Stout.

Ratings of brewery locations:

Here, we see a huge plot twist. We begin with breweries in Belgium being highest rated by far and four of the top five brewery locations are from Europe. Soon after, in 2008, breweries in the United States, California got the top spot. Furthermore, breweries in the United States, Oregon, and Norway join the race. After 2010, we see that breweries in Norway and Belgium gradually lost their prestige, and breweries in the United States took over the race completely such that in 2017 all the five top spots were occupied by the breweries in the U.S.

Popularity of brewery locations:

We begin with Belgium, England, Germany, and the Netherlands occupying the top spots, which is intuitive as these are the countries known for their beers. While the situation stays about the same until 2010, we see that breweries in the United States, California gain a lot of popularity. While breweries in Germany have lost some popularity in the last two years, breweries in England come with a final attack and finish the race in 1st place.

Brewing Bridges: Examining Beer Preferences Across Distances and Economies

Now that we have the beer preferences for each location and year, it is natural to be curious about how similar these preferences are, and how these similarities evolve over time. For this, we need to define a similarity metric. Remember that we have preferences in terms of beer styles and brewery locations and one metric for each of them is needed.

For beer styles, we have the top 3 styles for each location so we can use the Jaccard similarity metric. Then, the beer style preference similarity of the two locations is calculated between their corresponding beer style preference sets of size 3. For the brewery locations, we use the inverse of the geographical distance between the preferred brewery locations of a pair of locations to define how “similar” their preferred brewery locations are.

After calculating the similarity between each pair of locations for each year, we can take the mean of them to find the global similarity. Now, we can observe the trends of the means (global similarities). Below, we see two graphs showing beer style similarities and brewery location similarities, over time.

We observe some interesting trends in the data. It appears that beer style preferences have become globally more similar over the years when judged on ratings, especially since 2011. This means that the individuality of beer preferences is disappearing. Although the global similarity seemed to decrease during the last two years in terms of beer style preferences, it is difficult to discern if it is a real pattern or random fluctuations.

For brewery locations, it is a lot more volatile with many fluctuations. However, there seems to be a slight downward trend, especially for the more recent years. There are no other notable or obvious trends in the graphs.

However, this slight change in recent years has caught our attention: we decided to represent how the preferred brewery location evolved with the years for each country. In the visualization provided below, each arrow’s basis represents the brewery location and points toward the location in which this brewery location is preferred. Use the slider to change the year, and see how the arrow distribution changes over time!

It seems that with time, countries “un-cluster”, and countries prefer local breweries more and more, at least in terms of popularity metric. For example, for the majority of the states in the U.S, the most popular brewery location is California in 2006. As the years go by, most of these states prefer to consume more and more from their local breweries. These changes could potentially be attributed to the rise of microbreweries and local establishments, revolutionizing beer consumption patterns as highlighted in these two articles2 3.

In the face of a growing number of options from local breweries, the ever-increasing consumption of local beers becomes understandable. This phenomenon could contribute to a possible increase in divergence in brewery locations among countries.

As for the ratings, we can see that there is a much bigger interconnection between the countries, organized around a few poles. Therefore, reviewers seem to give more ratings to beers coming from their home countries, but not necessarily give higher ratings to those beers.

The role of distance and wealth between countries for beer preferences

One would logically think that there have to be underlying factors responsible for similarities and differences in beer preferences. In this study, we decided to focus on wealth and distance. For example, a relatively wealthy country like Switzerland might be willing to get more expensive beers than a much poorer country. Also, in terms of distance, we would expect that it is more likely that two states of the U.S. separated by 500 kilometers might have more similarities because of the interchange of culture and goods compared to similarities between a U.S. state and a country in Africa or Asia.

To determine how wealth and location played a role in beer preferences, we computed the 15 most similar and most different locations for those two facets. For location, we used the geographical distance between each location pair. For wealth, we used the difference in GDP per capita so that population and country size would not play a factor.

Then, using the pairwise preference similarities we computed before, for each location, we find the mean pairwise preference similarity of it with the 15 most similar locations to it and the 15 most different locations to it. Then, we aggregate these two similarity values over locations for each year. We also use the global mean preference similarity plotted before. Plotting all three curves in a single graph, we can visualize how “similar” and “different” locations behave in terms of their beer preferences with respect to each other and to the global mean similarity. As usual, we have four different metrics for each type of similarity: geographic distance and GDP per capita. Therefore, we end up with 8 different graphs and each one can be selectable using the dropdown menus.

As you can see, for each metric, the result seems to be the same: the “closest” group has always a higher similarity than the “farthest”, and the global preference is almost always between the two. Furthermore, the actual curves in the graphs seem to be the same in all cases except that the closest, global, and farthest curves are vertically shifted versions of each other. This indicates that both distance and wealth can be two of the many factors that affect beer preference similarity.

Interestingly, the curves are very close to each other for the wealth and distance graphs as well. One possible explanation is that the GDP per capita and the geographical distance are highly correlated since the locations in the same continent (individual states of the U.S., North European countries such as Norway and Finland, etc.) have relatively close levels of wealth and are also geographically close to each other.

When we perform t-tests and compare the resulting p-values for the 15 closest and 15 furthest locations for each similarity metric, we observe that in almost every case, the p-value was lower than 0.05. So we can say that the brewery location preference similarity significantly differs depending on whether it is calculated among close or far locations in terms of GDP per capita and terms of geographical distance.

Closeness in Terms of Beer Preference

Wouldn’t it be interesting to see which locations are the closest in terms of beer preferences? Of course, we have to keep in mind that we have four beer preference similarity metrics. To have a sensible single similarity value, we will find four beer preference similarities between each pair of locations, standardize each of the four metrics as z-scores, and then add them up into a single value. Then, the resulting overall similarities between each pair of locations for each year will be aggregated over the years such that we end up with a single beer preference similarity value for each pair of locations.

Now, we calculate the Pearson correlation coefficient of the beer preference similarity to either GDP per capita difference or geographical distance. The obtained values are -0.43 for the correlation coefficient between the distance and preference similarity, and -0.23 between the wealth preference similarity. Therefore, geographical closeness between countries seems to explain the similarities better than wealth. It is also important to note that distance between countries can be a confounder when we study the difference in wealth (if we take into account the wealth distribution in the world). Another hypothesis to explain those results is that the effect of wealth on beer preferences is U-shaped, as it is the case for the relation between income and beer consumption 4.

In any case, the correlation coefficients obtained are significantly negative. It means that the countries that are the closest in terms of distance and wealth are also more similar in terms of beer preferences. It looks like beer can give us information on the connection between countries! According to this result, one last question interested us: which countries are the most similar in terms of beer preferences?

Beer Brotherhood and 
 Enemyhood?

Beer is the perfect drink to exchange between friends to enjoy a relaxing night. One could even suggest that enjoying the same types of beer could be a great start to friendship! We call this amazing phenomenon “Beer brotherhood”!

However, we can observe this phenomenon not just between individuals, but between countries as well. It would be intriguing to see which locations are the closest in terms of beer preferences. For a single-value comparison, we use the aggregated z-score preference similarities mentioned above between each pair of locations.

Once we get these similarity metrics, we aggregate the similarity values over the years for each pair. Here, we regard the U.S. as a single country, rather than dividing it into individual states for a fair comparison. And
 Tadaa! We have the ultimate beer brothers! In the spirit of Lausanne - known as the Olympic Capital, we will now take them to the podium!

Are you ready to discover the top 3 beer brothers?

Drumroll, please!..

2nd

Belgium & Netherlands

1st

England & Scotland

3rd

Finland & Norway

As we can see, our metrics seem to allow us to capture cultural similarities. Indeed, the Beer Brothers are countries that have been historically linked together! England and Scotland are right next to each other and culturally similar. Finland and Norway are also both Nordic countries.

If countries can be Brothers through their beer preferences, they can also be enemies


Let us now look at the countries that would hate each other if they ever went to Satelite together!!

  1. Germany and Iceland
  2. Austria and Iceland
  3. Denmark and Iceland

We see one common denominator! It looks like Iceland has a lot of enemies to watch out for! And unsurprisingly, these countries are not as connected in terms of similar cultures so it makes sense that their beer preferences are quite different.

Final Comments

We are a group filled with beer enjoyers and beer-haters alike - We have our good and bad stories as you can see in our “small beer facts about us” section at the bottom of the page :)).

But ultimately, no one can deny the power of beer! And that is even more evident after going through the data and doing extensive analysis on the datasets.

It is a true sign of camaraderie when one can meet up with a friend, sip on some beer, and chat for hours. But each country, like individual people, has a distinct profile for the type of beer its people enjoy and consume. Looking at those preferences - how they change over time to become either more homogenous or distinct - leads to a lot of interesting conclusions. We see that people may potentially start liking the same types of beer more often. Furthermore, we see that both geographic location and wealth have a bearing on the similarity of preferences between two locations. And we can take those similarities to form a brotherhood and enemyhood of sorts, which results in conclusions that one would probably expect based on cultural similarities! So after lots of investigation, it appears that beer preferences are just one of the many facets that we can use to research the link between various locations. However, this dataset is definitely worth further investigation, and given more time, one can find even more fascinating links between beer preferences and the world around us.

References

  1. “Volume of World Beer Production”. European Beer Guide. Archived from the original on 28 October 2006. Retrieved 17 October 2006. ↩

  2. https://www.researchgate.net/publication/342563412How_and_why_did_craft_breweries‘revolutionise’_the_beer_market_The_case_of_Poland ↩

  3. https://www.mdpi.com/2306-5710/9/4/86 ↩

  4. https://www.researchgate.net/publication/281145522_Economic_Growth_Globalisation_and_Beer_Consumption ↩

Sip, savor, and meet the story brewers đŸ»

Yağız Gençer

Communication Systems

I once chugged a half-frozen beer, my sixth of the night in a total of 11, because my friends promised to buy the next round if I did. I got my brain half-frozen ever since.

Tuğba TĂŒmer

Computer Science

My favourite beer is Heineken. It's my favorite drink, I can have it every day and even drink 7 or 8 in one day. And yes, I'm proud of it.

Rishi Sinha

Communication Systems

This project started before I was legally allowed to drink beer.

Iris Toye

Neuro-X

I brew my own beer! And my favorite style is IPA.

Hadrien Sevel

Mechanical Engineering

I served a beer to a shirtless old guy at Montreux Jazz - turns out it was Iggy Pop.