51����

With the Cambridge-based programming learning platform Codio, they have analyzed students in a Stanford University programming course and sifted through the data to determine when any of those students are �spinning.� A student who is �spinning� is struggling on a programming problem and not making adequate progress toward the answer. For example, �it may look like theye working hard � but theye not working well,� said Dr. Sherman.

At this point in the project, theye able to identify a spinning student fairly easily, as well as chart how they come to their conclusion on the given problems. The next step is pulling out patterns and implementing an AI tech model to deduce those patterns for itself. If the model can start to pick up the patterns on its own for the given programming problems, it could potentially be applied to a broader range of problems and used to support students even further in the future.

�We�ve come a far way already,� Hinkle said, as the team prepares to make use of more of their data. They�ve compiled the work of 50 students into a data set to get the system up and running, but it only a piece of their six months of data they have � which is around 1,000 students � that they will continue to incorporate.

There hasn�t always been a clear solution or answer for the challenges of the research, but by applying the skills theye learning in the classroom, they�ve been able to tackle those questions and do their own problem solving with Dr. Sherman guidance.

But the best part of all the research so far? For Hans and Hinkle, it been actively pursuing and practicing their love of programming.

�We genuinely love it � the days of programming are genuinely the best days,� Hans said.