Trends — Alcohol Seasonality
Yearly Alcohol Calendar
People’s preferences for alcohol don’t stay the same throughout a year. It makes sense, it feels slightly weird to be at a 4th of July BBQ with a glass of cabernet. And there are holidays specifically centered around specific types of alcohol, New Years, Cinco de Mayo, etc.
Right off the bat we see a lot of things that make sense.
- The first and last days of the year see massive spikes in sparkling wine being delivered.
- We see that tequila has some local peaks around Cinco de Mayo (the biggest one), Fourth of July, Labor Day Weekend, etc.
- Red wine is much more popular during winter than it is during summer, and the Fourth of July is one of the lowest days we see for red wine.
- Rose dominates during summer, and sees its peak on the Fourth of July.
What we’re actually seeing here is the percentage share of each of these four types of alcohol, on a rolling 7-day basis. Around New Years, we see the share of deliveries for sparkling wine increase so much that everything else falls off the map.
Take a look at Cinco de Mayo as an example. Red wine didn’t actually fall off during that period of time, it’s just that the amount of tequila being delivered increased so much that the share of deliveries went heavily in the direction of tequila.
What You’re Hearing
You’re hearing maybe my all time favorite instrument that mankind has ever come up with, the mellotron. I use it all the time, they sound incredible, I wish I had a real one, but at the same time I don’t.
In this sonification, I’ve connected four (sampled) mellotron sounds to different alcohol types. The first thing you hear is a mellotron brass sound, representing sparkling wine. Then we hear a mellotron string section sound, tied to red wine. On Cinco de Mayo you can hear a brief blast of mellotron flute, and during the summer we hear a mellotron pipe organ take the lead.
Since mellotrons don’t have much control over dynamics, that’s where the data comes in. The share of alcohol deliveries controls a volume parameter on each track.
Here’s what all of the individual tracks sound like:
How it Works
I made a simplified tutorial on how to create these from start to finish. Check out the Colab notebook if you want to make one yourself.
Since we’re looking at the share of deliveries between these four categories, we would actually end up with a lot of quiet tracks if the 100% share was set to 127 and the 0% share was set to 0. No track would be able to reach the max volume unless all three other tracks were 0%. Everything was normalized so that the loudest peak was at 127 and 0 was 0.
After preparing the data, all we have to do is assign the normalized values to a CC track, one per alcohol type, choose an instrument for each type, and then assign that CC track to the volume. This one can be done with that sample notebook alone.
I see two main conclusions here:
- Mellotrons are the best
- People order different types of alcohol depending on holidays and the weather
The more interesting subtext here is that we’re looking at overall tendencies, not absolutes. I know there are plenty of people that will drink tequila all year long and don’t care about Cinco de Mayo. But think of this chart from the perspective of alcohol vendors, this is the reality of what they can expect to see shifting in their inventory throughout the year.
Postmates is always looking for creative data-focused people to join our team. If you want to make things like this, check out https://careers.postmates.com/ and say that Alex sent you.