Week 5

Peter and I have made good progress on our project at this point and are starting to get some preliminary results. We've been able to train predictive models based on information from the gradebook (such as lab, project, and test scores) and incorporating features from the Pytania data. By splitting up the Pytania responses into bins based on when they were answered we hope to capture a chronological picture of student engagement in the course. We also take the standard deviation of these bins to get a measure of consistency in engagement. Using this set of features to train a simple linear regression model we are able to achieve the results below when predicting final raw score in the course. Our next goal is to extract features from the zyBook data. We hope that engagement with the online textbook could indirectly measure student learning behavior's such as grit and dedication, and that these inclusions will improve our predictive accuracy.

Over the weekend my friend Matt came to visit. Sunday morning we watched the US women's soccer team thrash the Netherlands with a 2-0 victory in the world cup final!! Then we took the metro into DC and explored the National Mall in the afternoon. We went to the American history museum and the Hirshorn museum and later got rolled ice cream in Fairfax.