Do cold and hot cereals have statistically significantly different average ratings in the dataset?
Do cold and hot cereals have statistically significantly different average ratings in the dataset?: a task in SmolDataEnvs-harbor-train-sandbox: 📊 SmolDataEnvs: Harbor (train) (Harbor dataset). Files (in /home/user/input, no subfolders): - cereal.csv
The task
Files (in /home/user/input, no subfolders): - cereal.csv