Coffee Data Science
What Makes You Tick?
I was recently asked about my overarching goals for my experiments and research in espresso on the Decent Espresso Diaspora. I liked what I wrote, and I wanted to share those responses here with some slight editing. The questions are slightly modified to remove the previous discussion in the Diaspora.
Can you elaborate a little bit about your overarching goals for the research?
I’m working towards making the best espresso. That means tackling every variable, and often, it seems like it’s done in a wild, random pattern. Design of Experiment doesn’t mean you individually test each variable but rather you explore multiple variables at the same time. Over time, it is possible to tease each one’s influence out. With espresso, most variables are dependent on each other.
So it is a very interesting data problem as well which plays into my professional talents. I also have ADHD, so my process can seem from the outside a little wild and all over the place. In reality, I have multiple experimental threads going on at the same time, and I publish more thought out pieces as they develop without respect to a linear journey.
With respect to the best espresso, I don’t agree with industry. Most people making espresso aim for a 2:1 output to input ratio with the target of 18% to 22% Extraction Yield (EY). My target is a 1:1 with 22% EY. What started out as sifted staccato only, I’ve been able to learn from those shots to get better results for something unsifted.
For example, because I identified an issue with the water dispenser in the Decent Espresso machine, I tried a fix using a paper filter on top. That paper filter is also cut differently than others based on previous knowledge of how water travels horizontally in a paper filter.
For the roast I had been working with, it bumped my 1.3:1 shots to 24% EY, and my 1.1:1 shots went up to 20%. So my findings about the water dispenser suggests that everyone’s shots aren’t optimized because input water is not done well on the DE or other machines. That’s probably the main advantage of a lever machine with respect to water input.
How do I apply this information to the real-world?
I have a variety of articles that should at the very least inspire a person someone to experiment. Maybe you haven’t seen my medium articles, but if there isn’t something in the last 10 articles that can’t help you with a real world application, I would be surprised.
What is your North Star?
The best espresso. I’m obsessing over details and demanding high performance. I’m not concerned with making money from coffee. Coffee will never pay more than my day job. I want to taste a melted chocolate bar when I sip espresso.
Additionally, when I first introduced staccato, I was mocked, hated on, and received huge criticisms both publicly and privately. As I dove into any data available on coffee, I found there wasn’t much. Every time I challenged conventions, I was told I was wrong by people who didn’t believe evidence.
Like fines migration; no amount of evidence or experimentation would convince the people, who told me I was wrong, to admit that the evidence against the theory was compelling. Generally, I’ve only heard the sound of silence when I’m right.
I’m not going to wait for someone else to tell me how to make espresso better; I’ll do it myself.
If you like, follow me on Twitter, YouTube, and Instagram where I post videos of espresso shots on different machines and espresso related stuff. You can also find me on LinkedIn. You can also follow me on Medium and Subscribe.
