Another valuable lesson was learnt this week when using AI to report on xAPI Data. Looking back at this article, Using AI to report on xAPI – an Experiment, I had mentioned I had an LRS with over 4 Million Statements. One of my goals was to review data for a major spike in course completion in January 2024. This equated to around 1.8Million Statements.
For context, the courses were built using Articulate StoryLine and simply connected to an LRS to send the xAPI Data when packaged and published as SCORM. These courses were built in November 2023 and there was no custom xAPI included.
Recently, I setup the customGPT to accept xAPI data and so far the results have been pretty impressive. Using a slightly modified code base from Using AI to report on xAPI – an Experiment, I exported the data as CSV and proceeded to upload to the customGPT. First thing I was impressed with was that it excepted a 300Mb file with 1.8Million rows!
The customGPT recognised the data as xAPI (albeit cleansed as CSV) and started summarising. It was here that things started to unfold.
What I found was that the Object Names (object > definition > name) were not intuitive. What I mean by this is that Questions with verbs of answered etc were sending descriptions like ‘Pick one’ or Pick Many’. This was all set by StoryLine. The IRI’s for the object were also really random and required some much deeper diving to understand what they were.
I also notice that the grouping and parent in the xAPI statement were linked to the entire project, as opposed to a topic in StoryLine. This would help in drilling down the data as well as providing more conciuse reporting.
Answers [choices] to questions were also not intuitive and required additional work and linking to figure these out to get any real analytics from them.
What does this all mean? PLAN YOUR DATA FIRST!
Understand what it is you want to get out of your xAPI data. When developing the new courses, we are now placing more custom xAPI in our StoryLine files to better understand learn engagement and course performance. Take your time in planning this and you’ll be rewarded with much better outcomes.
Prepare and Use an xAPI Data Dictionary
Document what it is that you want to use, document the verbs and the objects. See what limitations you may have with the authoring tools as you may not be able to get exactly what you want.