An Exploration of Arabica Coffee and its Attributes

INFO 2950 Project

This research intends to investigate the connections between various coffee bean qualities, concentrating on fragrance, flavor, and acidity. The study addresses two main questions: (1) How well does the flavor of coffee explain or predict how aromatic the beans will be, and (2) How does the acidity in coffee compare between beans grown in Africa and those grown in Asia. It does this by using a dataset from the CORGIS Dataset Project by Sam Donald. The collection includes observations about the Arabica and Robusta varieties of coffee beans from various nations, together with ratings for characteristics like acidity, sweetness, and smell. The results of a linear regression study show a correlation between taste and aroma scores, with an R-squared value of 0.54 indicating that flavor accounts for 54% of the variance in aroma. According to bootstrapped hypothesis testing, the average scent score is predicted to increase by 0.63 to 0.707 points with 90% confidence for every additional point in the flavor score. According to a different hypothesis test, there is a significant difference in the mean acidity scores of the beans from the African and Asian regions, with the average acidity of African beans being a little higher. These results have ramifications for coffee growers, dealers, and consumers, indicating that fragrance can be enhanced by altering taste profiles and that regional variations in acidity should be taken into account when obtaining beans. The project acknowledges some limitations, including unequal numbers of observations for each region and potential biases in coffee scoring.