Learning Happens Everywhere
We’re no longer bound to one system, one platform, or one place for learning. These days, learners are watching videos on mobile, doing simulations in VR, getting feedback from peers, completing compliance modules, and reflecting in their digital portfolios — all on different platforms.
This is decentralised learning. And while it’s powerful, it can get messy.
To make sense of it all, we need a common language across tools. That’s where the xAPI Data Dictionary comes in.
What Is an xAPI Data Dictionary?
If xAPI is the standard that helps us track learning experiences anywhere, then a data dictionary is the rulebook that makes your data consistent, usable, and meaningful.
It’s a central reference that defines:
- The verbs you use (e.g. completed, rated, commented)
- The activities being tracked (videos, lessons, simulations)
- The extensions you include (device type, timestamp, assessment ID)
- How each xAPI statement is structured
What if every system uses a different verb for the same action? What if every course label labels the activity inconsistently? Good luck trying to run meaningful analytics on that!
Why It’s Crucial for Decentralised Learning
Let’s say your organisation uses:
- Moodle for core content delivery
- H5P for interactive elements
- A custom video platform for skills demonstration
- A third-party mobile app for on-the-job assessments
All of them support xAPI. But if you don’t define what verbs, activities, and context each one should use, you’ll end up with a data swamp.
With a shared data dictionary, you get:
- Consistent data across systems
- Cleaner analytics and dashboards
- Better insights into learner behaviour
- Future-proof interoperability
- Machine-readable data for AI and automation
Why ActivityID matters
Imagine a learner watches a training video, uploads a demo, and receives feedback — all through different systems.
If each xAPI statement references the same or logically structured activityId, your Learning Record Store (LRS) can stitch those actions into a coherent story of the users learning or assessment.
| Time | Actor | Verb | Activity (activityId) | Result/Context |
| 09:01 | Jon | watched | https://xapi.com.au/activities/forklift-video | Viewed full 5-minute safety demo |
| 09:10 | Jon | uploaded | https://xapi.com.au/activities/forklift-assessment | Uploaded 1:10 video of practical task |
| 09:30 | Reviewer | commented | https://xapi.com.au/activities/forklift-assessment | Timestamp: 00:48, Note: “Check shoulder check timing.” |
| 09:35 | Reviewer | rated | https://xapi.com.au/activities/forklift-assessment | Score: 4/5, Criteria met: 3/3 |
| 09:40 | Jon | acknowledged | https://xapi.com.au/activities/forklift-assessment/feedback | Marked feedback as read |
That activityId becomes the anchor — the common thread linking events, even across platforms. Documenting these relationships in your xAPI Data Dictionary can be as simple as:
| Activity ID | Name | Type | Used In |
| https://xapi.com.au/activities/forklift-video | Forklift Safety Video | Video | LMS, mobile app |
| https://xapi.com.au/activities/forklift-assessment | Forklift Safety Task Submission | Assessment Task | LRS, Remote Reviewer |
| https://xapi.com.au/activities/forklift-assessment/feedback | Feedback on Assessment | Feedback Interaction | Assessor dashboard |
By treating the activityId as the spine of your learning story, you empower every tool in your ecosystem to understand what’s going on — not just collect the data.
So now:
- Your analytics dashboards can report on complete learning journeys
- Your compliance tools can track who did what, where, and when
- Your AI models can make sense of patterns, outcomes, and feedback
Without consistent activityid, you’re flying blind — different systems may use different identifiers, and it all falls apart in the backend.
Verbs, URIs, and the Power of the ‘Did’
At the core of every xAPI statement is a powerful trio:
Actor – Verb – Object – Or more simply: Someone did something.
Use Standard Verbs Where You Can
Using verbs from official registries means your statements are:
- Easier to share across tools
- Instantly machine-readable
- Future-proofed for interoperability
Try starting with:
When You Need to Define Your Own
Sometimes the standard verbs just don’t cut it. You might need:
- endorsed – for managerial sign-off
- flagged – for issues raised during review
- validated – for peer-to-peer confirmations
In that case:
- Make your own URI (e.g. https://xapi.com.au/verbs/endorsed)
- Add it to your data dictionary with a clear definition and use case
- Use it consistently and document where it applies
- Ensure the URI resolves so a clear definition can be accessed
Your xAPI Data Dictionary becomes the shared truth about what each “did” really means.That little verb — the “did” — isn’t just a word. It’s a URI (Uniform Resource Identifier) that gives meaning and structure to the action.
Where to Start
Creating a data dictionary isn’t hard — but it does take planning. Here’s how to start your xAPI Data Dictionary:
- List all the platforms you’re using
- Identify key learning events and outcomes
- Standardise your verbs – Reuse known ones where you can and define custom verbs
- Define your activity IDs – Make them persistent and meaningful
- Map out context and result extensions – Things like roles, browsers, scores
- Document it all in a shared file – Start small, expand over time
Decentralised learning is here to stay. But without a shared structure for how learning data is recorded, we risk losing the very insights we’re working so hard to collect.
An xAPI Data Dictionary is the bridge that connects your tools, your data, and your impact across your decentalised learning environment.
So if you want your platforms to speak the same language — and your data to actually mean something — start with your dictionary. Because when the structure is sound, everything else can fall into place.
Need help creating your xAPI Data Dictionary or structuring your decentralised learning architecture?
Reach out at xapi.com.au or drop me a message on LinkedIn.
Written with some help from AI