Coffee Data Science
Coffee Color = Bean Temperature*
A strong correlation
Recently, I started doing analysis on roast variables to see what interesting aspects I could find. I became particularly interested with bean temperature or final bean temperature as it related to coffee color. One challenge in making experimental roasts is to make sure the roasts have similar colors, and if there was a variable that could improve this challenge, I would be very appreciative.
Bean color is highly correlated to final bean temperature in the initial data that I produced across 158 roasts. There is still some noise, but some of that noise is from different bean origins. For the same bean, is there a more precise correlation between temperature and color?
To test this question, I used a test coffee, and I ran my usual roast:
Then I pulled a sample every 5 degrees starting at 180C.
For each sample, I measured the color using the DiFluid Omix.
I ended up with a nice image of the leftover grounds by the end of the testing.
Then I plotted the data, and there was a very good linear fit to the data.
Even the Standard Deviation of color had a good fit.
The beans looked very oily by the end:
This data suggests that if you make two roasts of the same bean and end their roasts at the same bean temperature, the color will be the same or very close. In my first test with this idea, I ended up with four roasts within 0.5 points in color.
I’m going to use this concept to have a tighter control on my roasting experiments in the future.
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.
Further readings of mine:
My Second Book: Advanced Espresso
My First Book: Engineering Better Espresso
