Beer-in-Hand Data Science

• Typical Friday afternoon office conversation
  • How do you architect the ideal beer flavor profile visualizer?
    • In particular, how do you represent “hoppy, for a Kölsch”?

• Does the style Kölsch even describe a well-defined beer sub-group?
  • Would it be reliably distinguishable from non-Kölschs in a blind taste test?

Beer == water + malted barley + hops + yeast +

sometimes other stuff like fruit

• We categorize it into different styles based on
  • Type and ratio of ingredients
  • How the beer is made (e.g., how long and at what temperature it's brewed)
  • Regional and historical differences

BUT

• How well do styles actually demarcate the beer landscape?
  • Is there actually more inter-style variation than intra-style variation?
  • Is there a more empiricially clean way to categorize beers into super-groups?

In other words, we're asking: are beer styles just a social construct?