The average professor teaching an introductory (or advanced) data science class talks too much. Stop lecturing. Flip the classroom. Help your students learn by doing.
Lecturing is a horrible way of getting information from your head to theirs. Whatever speed you lecture at will be wrong for a majority of your students. For some, it will be too fast. They will fall behind, lose motivation and give up. For others, it will be too slow. They already understand everything you are saying. You are boring them. Instead, give them a reading/video/whatever to do before class. Class time is too valuable for you to drone on while they read Facebook.
Good soccer coaches know that “The ball is the best teacher.” You don’t learn to play soccer by being lectured at. You learn with the ball at your feet. The same applies to data science. You learn with your hands on the keyboard. The laptop is the best teacher. Every minute that students are typing, with their hands on the keyboard and their eyes on the code, is a good minute.
Time is scarce. At Harvard, the typical class meets for 2.5 hours per week. Students report spending, on average, 6 hours per week on each course outside of class. So, around 30% of total student time is spent in class. The most effective use of that time is hands-on-the-keyboard working with data.