xAPI is an interesting prospect in online learning. It “almost” fixes the learning data analytics challenges that technicians faced with the ancient SCORM standard. It gives a very straightforward, customizable API standard that technicians adopt to the learning environment of an organization to create an ecosystem that allows multiple modes of learning and analysis of performance.

This was one of my early tries at analyzing video using xAPI. I created the below prototype which wraps a video into an xAPI based wrapper, so that I can capture how different users interact with the video. This uses the native xAPI video profile and can capture data such as initialization, playing, pausing, seeking, player interactions, completion, and termination.

The video used here is from Wikipedia - "To the Brink: JFK and the Cuban Missile Crisis" , was created by the National Archives and its John F. Kennedy Presidential Library and Museum and made possible in part by the Foundation for the National Archives, the John F. Kennedy Library Foundation and the generous support of Lead Sponsor AT&T. Special recognition to the Lawrence F. O'Brien Family. At the National Archives Lawrence F. O'Brien Gallery, October 12, 2012, to February 2, 2013. This video gives an overview of the exhibit, the materials included and some background about the crisis.

Open the app in a new tab

image
The login to the test xAPi video data collection app

 

 

Now, I needed a way to monitor these interactions. I could monitor a multitude of actions as above but wanted to keep the focus narrow and sought to just focus on where the learners are pausing and if I can identify some common data points.

Now, I needed a way to monitor these interactions. I could monitor a multitude of actions as above but wanted to keep the focus narrow and sought to just focus on where the learners are pausing and if I can identify some common data points. I created a dashboard to visualize only the “paused” event as you see below. The original data is removed as it had used sample data from my colleagues. The data that you see below are from people that have accessed the above video from around the world.

Open in a new tab

The above version uses a new Watershed LRS, so a lot of the original data is lost. From the original version, some insights were really interesting. For example, a common paused time was found to be around 4:50 where the narrator talked about a fact where he contradicted himself.

Getting this sort of data from a learners would be really beneficial to the L & D teams on how to design their learning content.