If we want models to adapt to different courses we need some course indicators so the machine learning algorithm is smart enough to differentiate the type of course they deal with.

Some ideas + notes taken from discussions with others:

• Course typology (new course setting/s)
• Number of teachers in a course
• Course's teachers activity
• Course potential cognitive depth
• Course potential social breath
• Course using resources or dynamic activities / activities that encourage interpersonal relations instead
• Messages between students in the same course
• Indicators to measure the course difficulty (based on the overall course completion rate maybe?)
• Is completion enabled