BostonEdTech Meet-up Nov. 16, 2017

Dr. Neil Heffernan (WPI) discussed the use of Educational Data Mining techniques in improving educational technology. Dr. Heffernan's ASSISTments system, and online intelligent tutoring system for middle-school mathematics that has been hosted by WPI for over a decade. Last year ASSISTments was used last year by over 50,000 last year to solve 12 million mathematics problems, providing a rich data source for understanding student performance. I will talk about a cool use of that data.

In particular, Dr. Heffernan discussed the 2017 ASSISTments Data Mining Competition, sponsored by this Spoke as part of the NSF's initiative to help spur progress and innovation in educational research using big data. Competitors used educational data from ASSISTments to make long-term predictions regarding STEM career entry from 7th-grade click-stream data.

More details about the competition:

Why make long-term predictions about students career choices? School are already using early warning systems to help prevent students from sometimes catastrophic decisions, like dropping out of school, but less work has focused on student's interest in important educational domains (like STEM). These systems could not only improve students preparation for long-term educational trajectories, but also help to target short-term interventions for students who are exhibiting indicators of weak motivation and interest. The results of this competition could help inform the design of systems that could help try to reignite student interest STEM.

Competition, which runs through December, is already underway, and over 40 individuals and teams have submitting cross-validated predictions! Successful entries will be invited to submit both to a conference workshop (EDM2018, in Buffalo, NY) and to a special issue of the Journal of Educational Data Mining. Students, professionals and any one else with interest in data mining are welcome to compete.

You can join in our pursuit to better understand the predictive powers of early STEM engagement! Apply your own cross-validated prediction models using your preferred data mining techniques. For more information, please visit our competition website!

The Competition and Dataset

1. Both the competition and the dataset are available for free upon agreement and compliance with privacy rules.
2. The dataset is available without joining the competition, although we strongly encourage everyone to participate.
3. The competition will conclude on December 1, 2017 at noon EST.

The meeting will be held in McGrawHill in Boston at 6pm on November 16th, 2017. Be sure to register here.