Educational institutions have a wealth of data around student demographics, admissions, academic performance and more. In addition to these structured data sources, they also have unstructured data from sources such as student activity on academic discussion forums, campus network access and ID card usage. All of these data sources can be brought together in an institutional data lake to predict and influence student behavior – including attendance patterns, academic performance and time to graduate.
In this next DSC webinar, two Pivotal data scientists will discuss a recent collaborative project with a top university, in which many data sources were used to build a 360-degree profile of student activity on campus and help predict student success. The session will also provide an overview of the data science pipelines that were developed for training and scoring multiple models in parallel, in-database. These pipelines are now being used to predict student metrics (such as GPA, course grade and time to graduate), and even as intervention tools to help prevent students from dropping out.
Regunathan Radhakrishnan, Principal Data Scientist — Pivotal
Srivatsan Ramanujam, Principal Data Scientist — Pivotal
Hosted by: Bill Vorhies, Editorial Director — Data Science Central