Education Analytics of the Stanford Online High School
I wrote a series of programs to scrape and analyze data from class recordings at the Stanford Online High School (OHS). OHS classes are held over a video conferencing program called Adobe Connect. Each class is recorded using Adobe Connect's built-in recording system. These recordings essentially consist of XML describing the flash animations that make up the class, not actual class behaviors. I started by making a system to parse that XML and extract usable data from it. I then performed a fairly in-depth analysis of that data, now with more than 130 pages of visualizations showing everything from student talk time, the impact of teacher behaviors on individual student talk times, topic modeling of the text chat, characterizations of student "talkativity", to the most popular text chat colors.