How Data Science Can Unlock Teaching & Learning at Scale
summary
Coursera uses data science on a 50 million learner platform to build concrete products: In-course Help serves nearly 100 million personalized interventions with an explore/exploit message-level model and a learner-level model and yields double-digit gains in progress and completion; a Student Support dashboard predicts course grades with course-specific and cold-start models plus human-readable feature perturbations to guide counselors; a skills graph and skill scores adapt Elo-style updates over 300k questions (5% instructor-tagged) to produce explainable, real-time skill estimates for millions of learners.
AI-written summary · about
chapters (5)
- 0:00Introduction and journey
- 4:29Personalized interventions
- 8:04Human support efficiency
- 11:08Skill scoring for careers
- 16:29Data science for equity
related talks
02mBuilding data science capacity in your org
20mData Science Education in 2022
12mLearning Science @ Scale
48mHuman data interactions: experiments in how humans analyze data
19mData science education as a public health intervention in E. Baltimore
36mCore of Data Science and Open Source Resources
46mData Science Training in Communities with Limited Technology Resources
17mTeaching Data Science in Adverse Circumstances: Posit Cloud and Quarto to the Rescue