[ARCHIVED] Reporting on Average course engagement (inc. Mobile)

[jago_brown]

I’ve developed a crude report to provide the average number of student (CanvasData) requests for a course per day (over a 3 month period in query below)

(Total student course requests / active student enrolments) / number of days in requests record set, that students can access the course i.e. the course is published...

Can anyone spot any issue with my query below?

This is intended to provide an accurate total number of requests for each course in a canvas term:

PS I hope to add more on the report, later

Jago

[jago_brown]

I have used the query above with a few modification to create what i’m calling for now “module metric two” (mm2) in the following monitoring report (table 1), where I have also included Web-Browser time and Grades taken from the API:

table 1

General Table 1 information (mm2, A,B,C,D)

Positive

1. Provides crude idea of average student engagement with a course
2. Includes Mobile (but not mobile refresh token  - last use)
3. Provides data on ALL courses in the Canvas term
4. Relatively quick to run  query as comparing timestamp_month rather than full timestamp

Negative

1. Every request is equally weighted (yet bookmarking a page is of less learning value than reading the feedback on a submission)
2. ? This is what the post above was hoping to find out from other Canvas users

Requests analysis (A in table 1 and the SQL from initial post)

Negative item one above is why this mm2 is crude - one thing I've done that helps a little, is filter out activity when staff ‘Acting’ as a student (WHERE….. r.real_user_id IS NULL)

Table 2 below,  is where I've counted distinct types of request e.g.:

GROUP BY r.web_application_controller, r.web_application_action, r.web_application_context_type

Then looked at the type of ‘URL’ to get a clearer idea of the student engagement

table 2

More about table 1

• Grades (mm1) relates to this post: https://community.canvaslms.com/thread/18467-a-rudimentary-system-for-importing-grades-into-sis 
• Student browser-time (mm3) relates to this post url: https://community.canvaslms.com/thread/20722-how-to-calculate-students-total-activity-time-spent-in-... 

The posts or others below relate generally. I skimmed through them, but they were useful as I then knew others on the planet have also been asking and trying to solve similar questions.

https://community.canvaslms.com/thread/3158

https://community.canvaslms.com/thread/23104-canvas-data-page-views

Jago

[jago_brown]

I think we have improved on the above.  We (Bloomsbury Institute) are now defining our two module ( canvas course) monitoring review (MMR) analyitcs  / metrics as:

(1) Average student Requests - per credit - per module 

Requests come from student web browsers and mobile app sessions and are logged in canvasdata‌.  These requests can be for pages, submissions, files, login ... * This metric is the average number of requests between the start of teaching and the modules' (Canvas course) re-submission due date. 

 (2) Average student web browser time per module  

For each student module (Canvas course) enrolment - Canvas keeps a running total of (web browser session) time spent.  This metric, sums and averages the time of all students at the re-submission due date. 

and have made some progress towards automating their calculation in a Python Django project

* More fully defining and classifying the different request types is a job to be done.  Have Instructure provided a taxonomy some where?

Jago