Exploring the Factors that Drive Coffee Ratings

Coffee is more than just a morning ritual; it is a cultural staple for many people around the world. One national report showed that 67% of American adults consistently drink coffee daily, a higher percentage than all other common beverages such as tap or bottled water. In this work, we delved into factors influencing coffee quality reviews, such as sweetness, flavor, aftertaste, bean location, and different coffee producers. Using publicly available data from the Coffee Quality Institute (CQI), we analyzed these influencing factors by implementing a regression model via gradient descent with the coffee rating being the attribute of interest to predict (i.e., the target attribute). We hypothesized that both the coffee producer and sweetness are the two most important factors influencing coffee ratings. Our analysis shows that a linear regression model is well suited to analyze this data and as a result, only a few factors contribute significantly to the prediction of coffee ratings. In particular, we find that sweetness is by far the most important factor to predict total cup points, a comprehensive coffee rating based on several sensory and production related attributes, with other sensory related attributes also playing a key role in our analysis. Surprisingly, we see that the coffee producer has very little predictive power in our model. These results give insight into consumer coffee preferences, which is valuable information for manufacturers globally.