The popularity of beer has exploded in the last decade, with the number of available craft beers increasing by over 1000% in the last decade. There are now over 80,000 hoppy beers being made in the US alone! With this fervor has come thousands individuals, both casual and experienced alike, who actively seek out and review every beer they can find. I use 'big data' tools to aggregate these vast swaths of information industry and consumer behavior, often resulting in data sets that consist of million individual observations. This allows me to make insights into how individual and population preferences change over time and in response to environmental information.
Documenting the hipster effect
A common stereotype is that 'hipsters,' or really any expert for that matter, tend to not things that are popular. To investigate this I modeled how individuals score a beer in response to both their own experience level, and the popularity of the beer. If the hipster effect is real, then experienced reviews should dislike beers that have a large number of reviews.
Indeed, we find exactly that. The figure at the right shows how the scores of individual reviewers who are inexperienced or moderately experienced (blue and black likes) respond to popularity. The flat lines indicate that they are indifferent to popularity, rating popular beers the same as less-popular beers. Yet, the red line, which represents the experienced reviewers, indicates that they score popular beers much lower than less-popular beers.