I have developed an open learning analytics system that will track student interaction data at the “activity level” or the “transaction level”, depending on what you call it. This system is currently composed of a quiz web application, a video web application, and a dashboard. The quiz application is our instance of the open source quiz app Open Assessments from the Lumen Learning github. The video application (it actually does way more than just track video analytics) was developed at Brigham Young University and is called Ayamel. Here is a diagram of our system:
To explain this system, let’s start at the Learning Management System (LMS). The student logs into the LMS and sees a bunch of links. Some links go to videos, some links go to quizzes, and some links go to the visualization app. If they click on a link, they will be authenticated into the app (so we can track who they are for analytics purpose) and are taken directly to the resource indicated in the link. Technically, this is an LTI launch (which is more than just a link) but we don’t have to go into that here. Once the student is in the app, xAPI statements are sent from the app to our Learning Record Store (LRS). We are using Learning Locker (an open source LRS) for our LRS. This LRS will store all of the statements that are sent to it from our apps. Then, our visualization app will take the data in the LRS and will display it back to students in real-time. We can do this because we have real-time access to the data in our LRS.
The data that we are tracking in this system includes the following:
|Video Analytics||Quiz Analytics||Visualization Analytics|
|Number of plays||Number of question attempts||Number of visualization views|
|Number of pauses||Time spent on quizzes||Time spent in visualization|
|Number of skip ahead/back||Number of quizzes attempted||Number of video links clicks|
|Number of speed changes||Average confidence level||Number of quiz links clicks|
|Average video watching speed||Max number of attempts|
|Number of volume changes||Max time on a quiz|
|Average volume setting|
|Number of mute/unmute|
|Number of max/minimize|
|Time spent on videos|
|Number of videos watched|
We are currently collecting data and have more than 200,000 statements in our LRS.