Mismatch between API data and New Analytics

[MichaelLangUG]

Can somebody please answer this question for me.  I am experiencing very confusing anomalies between data obtained via API calls and corresponding data obtained via the New Analytics UI.

Specifically, I ran the following API call to extract details of student activity on a course:

https://<canvas>/api/v1/courses/<course_id>/analytics/users/<user_id>/activity

However, the values retrieved in this way do not match those obtained via:

New Analytics > Students

Any ideas what is going on?

I ran a scatterplot of two sets of values from my class of 217 students and there is absolutely no correlation whatsoever between the data obtained via API call and the data obtained via New Analytics download.

[stimme]

The API endpoint returns live data. New Analytics uses snapshot data. New Analytics used to display a timestamp for the data snapshot, but that was removed as of February 14. This might explain the differences you are seeing.

[MichaelLangUG]

Thanks.   This not as simple as live data versus snapshot data.  The discrepancies are massive, and when looking at data from several months back, there is no excuse for such huge variances between New Analytics and the API data.

To give you an indication of just how much out of line the two data sets are, here is a scatterplot. It should be a straight line but it's a total shotgun map of data.

API data correlated against New Analytics data (total page views per student, standardised percentiles):

API data correlated against New Analytics data (total page views by each student)