Updated all pages and added new plots

epfl-ada · Dec 22, 2023 · b3c982d · b3c982d
1 parent 8e17340
commit b3c982d
Show file tree

Hide file tree

Showing 17 changed files with 81 additions and 22 deletions.
diff --git a/Beer styles.md b/Beer styles.md
@@ -3,11 +3,18 @@ layout: default
 ---
 
 # Beer styles
-
 There are many specific beer styles, which have different alcohol content!
 
 Let's group them by general style and explore the ABV and users' preference correlation!
 
-![Avg_ABV](./graphics/plots/style_abv.png)
+![Avg_ABV](./plots/style_abv.png)
+_We see that some beer styles are significantly stronger than others. If these styles are consumed more frequently on one of the two continents compared to the other, then this could explain some differences in the way ratings are correlated with ABV._
+
+![Avg_ABV](./plots/style_count.png)
+_Americans are dominating Europeans in the number of ratings per beer style... but..._
+
+![Avg_ABV](./plots/style_rating.png)
+_nevertheless, the average rating per beer style between the two groups is comparable. American ratings are always higher, except when it comes to Porters! It looks like Europeans like this beer as much as Americans do!_
 
-![Avg_ABV](./graphics/plots/style_rating.png)
+### Conclusion
+Americans are huge fans of beer styles with high alcohol content (above 6%), such as **Stouts, Strong Ales, India Pale Ale, Porters, Bocks** and **Dark Ales**. We can see this from the number of ratings, as well as the high rating average. **Strong Ales, Stouts** and **India Pale Ale** have an average rating greater than 3.5 and the number of ratings for these beer styles is double, or even 4-5 times grater than the number of ratings Europeans submited.
diff --git a/Data sources.md b/Data sources.md
@@ -3,14 +3,11 @@ layout: default
 ---
 
 # Data source
-
 The data we are working with was gathered from the following two websites about beer lovers:
-
 - [RateBeer](https://www.ratebeer.com/)
 - [BeerAdvocate](https://www.beeradvocate.com/)
 
-#### From each website we used several datasets
-
+#### From each website we used several datasets:
 - **users** (_American and European users only!_)
 - **ratings** (_`user_id`, `date`, `abv`, `rating`, `appearance`, `aroma`, `palate`, `taste`, `overall`_)
 - **styles_matching** (_Generalising the `specific` beer style to its `general` style_)

diff --git a/Naive analysis.md b/Naive analysis.md
@@ -3,23 +3,27 @@ layout: default
 ---
 
 # Naive analysis
-
 Moving straightforward to our research question:
 
 **Do Americans prefer beers with higher alcohol content (ABV) than Europeans?**
 
 --->We investigate the average rating that EU and NA users give, for each ABV rounded percentage.
 
-![Avg_ABV](./graphics/plots/Avg_ABV.png)
+![Avg_ABV](./plots/Avg_ABV.png)
 
 - _It seems like NA tend to give better ratings in general_
 - _All ratings are higher for beers with higher ABV_
 
-Does the number of ratings per rounded ABV percentage has to do something with the average rating per ABV?
-#### **TODO: decide wether we plot both separately or merged, and if we cut the ABV to 20%**
+This plot gives some intuition for the answer, but in order to be sure, we must perform further statistical anaylsis -> [Statistical analysis](Statistical%20%20analysis.md)
+
+We also want to investigate if the number of ratings per rounded ABV percentage has to do something with the average rating per ABV?
+
+![Avg_ABV](./plots/count_abv.png)
+
 _Beers with ABV lower than 5% or greater than 15% have fewer number of ratings. If only few users rated, and graded them good, this results in better average rating for the specific ABV._
 
-Is there a correlation between preference and a specific beer style, known for higher ABV?
+Another question arises out of curiosity:
+**Is there a correlation between preference and a specific beer style known for higher ABV?**
 
 Let's investigate this question:
-[Beer styles](/ada-welovepandas-webpage/Beer%20styles)
+[Beer styles](Beer%20styles.md)
diff --git a/README.md b/README.md
@@ -1,20 +1,17 @@
 # Beers Across the Atlantic
-
 ## Initiative
-
 **Beer** :beer:
 the World’s Most Beloved Drink
 
-It is accepted worldwide that **beer** is the "most beloved drink", but who is leading the competition?
-<br>
-We would say Europe and America are the biggest beer consumers among all continents.
+It is accepted worldwide that **beer** is the "most beloved drink", but who is leading the competition? 
+</br>We would say Europe and America are the biggest beer consumers among all continents.
 
 But again, do Americans and Europeans drink the same beer types, or is there a beer attribute that makes a difference between this two big consumer groups?
-
-### Let's investigate the beer preferences across the Atlantic
+#### Let's investigate the beer preferences accross the Atlantic!
 
 ## Description
-
-Using a large dataset, which provides beer ratings and reviews on beers from people all around the world, we will investigate the difference in beer preferences across Americans and Europeans, and try to find the reason behind it! Embracing the 14 beer general styles provided, which differ in characteristics as ABV, appearance, aroma, palate and taste, we expect to find interesting results.
+Using a larga dataset, which provides beer ratings and reviews on beers from people all around the world, we will investigate the difference in beer preferences accross Americans and Europeans, and try to find the reason behind it! Embracing the 14 beer general styles provided, which differ in characteristics as ABV, appearance, aroma, palate and taste, we expect to find interesting results. 
 
 We suspect that cultural dissimilarities among this two groups might lead to different inclinations when it comes to alcohol concentration in beers. Moreover, we will inspect the correlation of beer type with this inclination and we will deep into the evolution of the beer preferences between 2000 and 2017, hoping to find some difference among America and Europe.
+
+
diff --git a/Seasonality.md b/Seasonality.md
@@ -0,0 +1,20 @@
+---
+layout: default
+---
+
+# Seasonality
+### You must have heard of Oktoberfest!!!
+
+We suspect that the difference in the trendline of American and European ratings is due to this big event taking place in late September and the first weekend of October. Oktoberfest, is in fact the largest beer festival in the world!
+
+In order to investigate this speculation, we will first decompose the time series for each region into seasonality and trend:
+
+![Avg_ABV](./plots/seasonality_na_vs_eu.png)
+_North America has a clear seasonality pattern, but Europe shows two main peaks each year. If one of them is Oktoberfest, as we supposed, what could be the second one??_
+
+Do you know this guy?
+![Avg_ABV](./gifs/stpatrick.gif) 
+**TODO: make image smaller**
+
+Have you ever heard of St. Patrick's Day? 
+- It is an Irish traditional holiday, taking place in March.
diff --git a/Statistical analysis.md b/Statistical analysis.md
@@ -0,0 +1,5 @@
+---
+layout: default
+---
+
+**TODO: add the statistical analysis**
diff --git a/Time evolution.md b/Time evolution.md
@@ -0,0 +1,23 @@
+---
+layout: default
+---
+
+# Evolution over time
+Diving deeper and deeper into this data, made us question if the tendency for beers with higher ABV has increased over time. 
+
+Do Amricans or Europeans like beers with more alcohol as time goes by?
+
+Let's take a look at this plot:
+
+![Avg_ABV](./plots/abv_time.png) 
+_It looks like it is true! There is a slight evolution of the ABV over time!_
+</br>_The graph clearly shows that gerenally North American consumers drink stronger beers on average than European._
+
+That's not all! We discovered another very interesting thing!
+</br>Look at the shape of the trendline! It is periodical!
+
+This gives an insight for the seasonality of this magical drink. Even though it is more noticable for North Americans, ABV has peaks every start of a new year and a low peak each summer, on both continents. But why is the shape of the trendline different for Europeans? Why is it not that clear, as the one for North Americans?
+
+If you want to know more -> take a look at the [Seasonality](Seasonality.md)
+
+`Note 1: The fluctuations present in the first few years are not relevant for observation. It is possible that the reason behind this noise is the very few users of both BeerAdvocate and RateBeer at the begining of their existance. As the popularity of this websites grew, and more users started to register, the curve started to take its representative shape.`
diff --git a/_config.yml b/_config.yml
@@ -15,8 +15,14 @@ navigation:
     url: '/Data sources'
   - title: Naive analysis
     url: '/Naive analysis'
+  - title: Seasonality analysis
+    url: '/Seasonality analysis'
   - title: Beer styles
     url: '/Beer styles'
+  - title: Time evolution
+    url: '/Time evolution'
+  - title: Seasonality
+    url: '/Seasonality'
 
 # Exclusions
 exclude:

diff --git a/graphics/gifs/beer_loading_gif.gif b/graphics/gifs/beer_loading_gif.gif
diff --git a/graphics/plots/Avg_ABV.png → plots/Avg_ABV.png b/graphics/plots/Avg_ABV.png → plots/Avg_ABV.png
diff --git a/plots/abv_time.png b/plots/abv_time.png
diff --git a/plots/count_abv.png b/plots/count_abv.png
diff --git a/plots/rating_time.png b/plots/rating_time.png
diff --git a/plots/seasonality_na_vs_eu.png b/plots/seasonality_na_vs_eu.png
diff --git a/plots/style_abv.png b/plots/style_abv.png
diff --git a/plots/style_count.png b/plots/style_count.png
diff --git a/plots/style_rating.png b/plots/style_rating.png