As Machine Learning moves past the peak of inflated expectations, I am getting involved in dragging it towards the slope of enlightenment.
Recently I’ve been doing a lot more concept work to integrate Machine Learning into our product’s repertoire. This is all related to the learning analytics work we’re doing, and the sheer volume of data we’re expecting to start flooding in.
However, when I’m just picking through the masses of data we get per. partner, per module, per. term… it’s can become overwhelming just working out what questions we want to ask, what theories we want to test? Having the data and no questions seems very much like I’m putting the horse before the cart.
However, here’s the first version of the big questions I want to ask. I think three questions to start is plenty for one team!
- What are the learning patterns of our adult learners (per. programme)
- Adult learners, are they working the nights, weekends or days?
- If we can make sense of when they are working, can we design better interventions?
- When is the lowest engagement per. module (per. programme)?
- Engagement is a lead indicator of learning achievement
- If we can find low engagement, can we identify and intervene
- Can we design differently for next time
- What are our highest engagement activities across all of Keypath?
- What leads to it, and can it be packaged and used elsewhere?