Relating Coffee Species to Brew Taste and Region with Machine Learning

If coffee flavor could be correlated to species and even country of origin, coffee purchasing would be much simpler, as consumers could use their own preferences to determine what to buy. This study aimed to find that correlation, using Data from the Coffee Quality Institute, taken from kaggle, and make predictions on the species based on the taste characteristics of the beans.

Buying coffee beans could be a confusing task, due to the large variety, each with their own region and flavor profiles. For many consumers this could lead to overbuying, or making unwanted purchases. If coffee flavor could be correlated to species and even country of origin, coffee purchasing would be much simpler, as consumers could use their own preferences to determine what to buy. This study aimed to find that correlation, using Data from the Coffee Quality Institute, taken from kaggle, and make predictions on the species based on the taste characteristics of the beans. Due to limitations with the data, and keeping in mind that the models tested weren’t tuned to perform their best, the Random Forest model performed best, having one of the highest accuracies with the most favorable set advantages and disadvantages.

Explore More!

Research Paper