Header Image

BTA


The Beer Trip Agency

Welcome to the Beer Trip Agency!

🍻 Greetings, beer lovers! 🌍✈️

Get ready for a frothy escapade with BTA - Beer Trip Agency, where your taste buds embark on a world tour, guided by insights from the renowned beer enthusiasts at RateBeer and BeerAdvocate!! πŸš€

🌐 Envision this : A journey that spans borders, cultures, and continents, all in the pursuit of the gold liquid – BEER! 🍺 Thanks to our collaboration with the legendary RateBeer and BeerAdvocate communities, our hop-centric expeditions will take you sipping saisons in Belgium, clinking steins in Germany, and grooving with IPAs in the USA. It's not just a trip; it's a taste odyssey, where each pour narrates a story, and every brewery is a chapter recommended by the discerning palates of RateBeer and BeerAdvocate.

🍻 BTA isn't just a travel agency; we're your companions in beerventures, dedicated to transforming every trip into a sudsy spectacle, fueled by top-notch recommendations and reviews from RateBeer and BeerAdvocate. Our specialty: designing your trip, and when we say your trip, we mean a trip fully organized based on your personal criteria! So, hop on board, grab your passport, your pint glass and let's toast to the world of beer exploration, guided by the knowledge of RateBeer and BeerAdvocate! These two remarkable websites have earned our unwavering trust, and, based on our analysis, we hold them in equally high regard, a sentiment reflected in the compelling graphs presented below:


CDF of standardised user scores from Rate Beer and Beer Advocate


This plot shows the CDF of each website. By simply looking at it, we can make the hypothesis that BeerAdvocate users tend to give higher z_score than RateBeer users. We will test this hypothesis later on.


Scatterplots of number of ratings vs scores for both websites (left - Beer Advocate, right - Rate Beer)


These two plots give us a way to visualize the zscore distribution of the 1% of the most reviewed beers from both sites. It also gives us an idea of the number of times that the most-rated beers are reviewed.

After graphing the data and analyzing it a little in the previous graphs, certain hypotheses need to be tested so that they can be used correctly in the future. Indeed we are trying different tests to see if the z_score distribution of the beers from the two websites are directly comparable.

The CDF plot gave us the intuition that the z_score distribution of the beers from the two websites is not directly comparable. We tried to confirm this intuition by doing two different hypothesis tests that we thought were well suited to test our hypothesis. We first computed a permutation test which randomly shuffled the data from the two different datasets and evaluate if the difference of z_score distribution is significant. The test gave us a p-value of 0.0 which confirms our initial intuition that there is a significant difference between the two distributions. To be sure, we also tried to do a Kolmogorov-Smirnov test which gave us a p-value of 1.16e-24 which also confirmed our intuition.


Comparing review numbers between the sites (left) and unique beers between the sites (right)

The first plot shows the difference in the number of reviews year by year of the two websites. We can see that while the number of reviews of both sites is increasing before 2014-2015, BeerAdvocate tend to grow faster than (or has more active users) than RateBeer. We can also observe that both sites decreased in the number of reviews after 2015.
The number of reviews in 2017 for both sites is particularly low, this could be due to the fact that the dataset may have been collected during 2017 and that the number of reviews of 2017 is not complete.

The second plot shows the proportion of beer present on both websites versus the proportion of beer present on only one of the two websites. We can see that the majority of the beer is present on both websites, which allows us to work directly with matched files without "losing" too much data.




Wow, we already have some customers interested! Nice to meet you Bob, Beatriz, Akhil, Manouel, Marija, Martin, Tim and Mohammad. Tell us about your story… What are you doing in your life? Oh, we see, so you are responsible for training tomorrow's data scientists? And what exactly does it involve? Wait, WHAAAAT ?! 122 projects to correct! Oh my god, you guys must be exhausted after such hard work, you've probably been wanting to take a break and travel for a while.

All right, buckle up as we dive intoindividual preferences!Let's take it step by step and discover what each beer enthusiast savors on their hop-filled journey!!