What percentage of the dataset is composed of spam messages based on the initial analysis?

What percentage of the dataset is composed of spam messages based on the initial analysis?: a task in SmolDataEnvs-harbor-train-sandbox: 📊 SmolDataEnvs: Harbor (train) (Harbor dataset). Files (in /home/user/input, no subfolders): - spam.csv