Beef Quality Classification and Logistics Patterns over Local, Regional, and National Supply Chains in South Korea

In this paper, we intend to examine the impact of several biological factors such as cattle types, gender, and weights on beef quality using data sets obtained from the digital beef traceability system in South Korea. In addition, we also like to identify several cities or slaughter houses that produce the highest quality of beef to validate whether best places for Hanwoo in our analysis are consistent with cities constantly recommended from Google query and many Koreans. Then we calibrate a machine learning model to identify cattle that are most likely to produce the highest-grade beef. According to our calibrated decision tree (DT) model, neutralized male Hanwoo with a certain weight range is most likely to produce high grade meat and our DT model performs at least twice better than a random model in terms of correctly identifying positive samples when top 20% and 40 % of cattle were chosen for a prediction task. Finally, we like to explore different types of beef supply chains in the current digital traceability system and profile them based on their distinct geographical coverages and beef consumption patterns.