Density — Postmate Density

February 3, 2020

Share on FacebookShare on TwitterShare on LinkedIn

Density — Postmate Density

Click here for more info on this project.

Cities

It’s not news that no two cities are the same. There are so many different ways to compare cities, walkability, public transportation, food, culture, demographics, size, density, history, etc etc. We have a lot of data on a lot of those different things, but we’re going to focus on size and density right now. Let’s just look at San Francisco and Miami.

Size is pretty easy, we’ve got coverage zones for every market, so it’s just a matter of finding the square mileage of those zones.

Density is a bit trickier, but maybe more fun and interesting. There are a lot of angles at which we can look at density, how packed together are our customers, our merchants, or our Postmates for example. In this situation I decided to look at how many Postmates are on deliveries in each market, and I wanted to see how it changes throughout the day.

This is what the total number of active Postmates looks like for San Francisco and Miami.

And this is what it looks like relative to the physical size of the market, ie: how densely packed are our Postmates in the market as the day goes on.

I fully admit that I cherry picked these markets to make the point that density and total number of Postmates mean two very different things, and can help describe the characteristics of a city.

What You’re Hearing

I made the exact same synth patch play two simple arpeggios, one for each market.

The more total Postmates that are on active deliveries, the louder the sound becomes. The more dense the distribution of Postmates is, the more distorted the sound becomes.

Distortion is kind of an unfair effect because it compresses the sound as it increases, which can give it the illusion of higher volume. If you listen to the highest peak during dinner, the Miami track actually is much louder than the San Francisco track, despite how completely fuzzed out San Francisco sounds (if you’ve got volume meters, you can see how much higher the level is on the Miami side).

Our Miami zone is quite big, sure there are areas of Miami that are dense, but overall, things are a lot more spread out. So we don’t really hear that much fuzz on the Miami side. But overall we do have a lot more active Postmates during the course of a day.

Our San Francisco zone is pretty much just the 49 square miles of the city of San Francisco itself. When it’s the peak dinner rush in SF, the odds of a Postmate running into multiple other Postmates at some point during their day are high.

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.

This was another relatively simple one, we’re just controlling parameters of the sound with the data, so we just need two Control Change (CC) channels per MIDI track. In this case we’re generating one MIDI track per market, and each one has one CC channel for the total number of Postmates, and one CC channel for the Postmates per square mile. Both markets are using the same scale, set to the min and max of either market for each metric.

Once we’ve got our MIDI tracks ready to go, I dropped them into my favorite Digital Audio Workstation, REAPER, and loaded up Waves Element 2.0, dialed in a sound to be used on both tracks, set up a different arpeggio to constantly repeat on each, added in a fuzz plugin, then mapped the CC tracks to the volume of the synth, and the fuzz amount on the fuzz plugin.

The data performed the rest.

Conclusion

Every market we have is going to have their own signature when we look at two different metrics like this. Some of our markets have exponentially more active couriers than others, so it would be hard to compare them all on the same scale like this. Imagine a graph looking at the number of cars that cross the Bay Bridge in the course of a day vs the number of cars that drive across the Arizona-New Mexico border in I-10 in the course of a day. You wouldn’t even be able to see any shape at all from the latter if they were on the same scale.

But when we look at markets that are pretty close together in size, we can really focus on how the unique flows of Postmates through the minutes in a day. In my mind it makes it feel like each market is its own organism as it inhales and exhales throughout the day. One market might be absolutely massive, but feel more sparse and orderly, while another could be relatively small, but be completely packed and frenetic. Or It could be huge and dense and small and sparse, or somewhere right in-between, we’ve got examples of everything.

The big takeaway here is that being a Postmate can be a vastly different experience in different markets. And we have to account for these differences in every way possible to make sure that the experience is still up to our high standards everywhere.

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.

Engineering

More from Engineering

View All

Density — Market Drum Machine

After making the atonal cacophony that was the Regions piece [link to it], I decided to try to set things up in a way that would result in something a bit more accessible. An electronic drum kit seemed like a pretty good place to start.

February 3, 2020