experience and targeted interven,ons to help students – Create insights into the learning challenges students face and how interac,ons with their learning environment affect them – Create insights into the course and program content efficacy Significant progress: – State-‐of-‐art data plaKorm in place (hybrid data warehouse and big data plaKorm) – Apollo Classroom connected to the big data plaKorm – Data-‐Driven Applica,ons, such as risk-‐alerts and personalized study guide, rolled out
to: – Enable large-‐scale learning analy,cs – Build predic,ve models by analyzing student behavior and measuring learning • Strengths of this approach – Scalability at commodity compu,ng prices – Flexibility to handle structured and unstructured data – Reusable infrastructure for all future applica,ons
learning objec,ve Direct links to instruc,onal resource (textbook readings, videos, other media) Contact from an academic counselor Synchronous resources: -‐ Faculty-‐led sessions -‐ Live 1:1 tutoring -‐ Live labs Struggle with a learning objec,ve Counselor ,p about a job lead Life circumstance Quiz results Course ac,vity Phoenix Career Services inquiry Assignment grade Journey toward a degree Example Events Risks Interven,ons
inference engine allow us to build, test and deploy models predic,ng a student’s con,nued aVendance at the university, the key reten,on metric. • A variety of in-‐course student behaviors contribute to the predic,ve model. Data collected for associates, bachelors and graduate students Program Progress Number of courses passed in program Number of courses passed prior to program Level of activity Number of weeks of missed attendance Missed attendance in previous week Number of days since last post Number of days submitting discussion posts this week Total number of posts in the current week No assignments submitted in the most recent week Missed assignment, weighted by assignment size Number of days content was accessed in most recent week Mastery Current course points earned divided by attempted points Student data can power high-‐value predic,ve analy,cs that result in improved student outcomes.
all of the classroom discussions are about introduc,ons. These are essen,ally biographical entries. • We used some natural language processing to extract key terms about work, home, hobbies and school, etc. • From this we built a “social” model of the student based on their self proclaimed social aVributes • Next steps – Segment the student popula,on according to a social profile – Do an auto-‐cohor,ng experiments – Suggest “students who may have similar interests”
to decide on specific intervention As Next Step venture into prescriptive analytics One Challenge involves culture change for frontline staff One Baby Step try thumbs-up and thumbs- down
• Data Aggregation skills - Hadoop, Warehouse • Business rules behind data • Ensure High Data Quality Data Scientist • Build Features • Test Models Integration Engineers • Web Application Developers Manager • Remove Roadblocks • Strong business understanding