Contacting users based on course details

[jon-betts]

Tasks

• A basic version of this is now available on orgs
• Example org page here
• Example report here

Current work

• Individual reports are gathered here: https://report.hypothes.is/collection/76-teacher-activity
• Linked to from orgs: e.g. https://report.hypothes.is/dashboard/6-organization-focus?organization_id=us.lms.org.TARKQ36QQAGEfIowmGOuGQ&timescale=semester&role=user&group_type=course
• Linked to from team member pages: e.g. https://report.hypothes.is/dashboard/19-prototype-owner-focus-dashboard?owner_id=241592658&timescale=trailing_year&role=user&group_type=course

---

Some extra detail on potential requirements here:

• https://docs.google.com/spreadsheets/d/1e9DrQO8eqWsIBPXVsLNd5xG1Mk4pc07gWnWIVBjh5tM/edit#gid=0

This is still WIP

User story

As a success person I would like to conveniently contact teachers from a courses I pick from a list in an organization I choose.

Background

In Report a user can currently see:

• All teachers who are in an organization
• All teachers within a specific course

If you, for example, want to browse the courses and find all teachers who are in Spanish courses you identify by the name, then you have to click through to each course one by one and compile the names.

This is too much bother, so users rely on a report in the old reporting system which is powered by Jon U data. This means it’s less accurate, is not oriented by organization, and is often offline for portions of the day.

Solution suggested by user

Add comma delimited emails of teachers as an extra column to the group activity table.

Priority (Very High)

This is something being asked for a lot, and also provides a sub-optimal solution to many other use cases.

Effort Estimate (Medium)

We have all the data we need for this report, it’s just a case of arranging it differently. Assuming we solve this as suggested with a single report this should be pretty quick to turn around. This is slightly complicated by the fact that the user data must remain resident in the region (we can’t aggregate it in the US). This makes this a harder query than you might expect for a similar looking report from all US data. If as per Jeremy’s comment we require this over multiple organizations we may have to investigate this, as the current report is for a single organization.

There are some open questions about how we present the user data, as there are multiple users per course. I think there might be some work to do here to see if this is the optimal way of presenting this data. I believe the current report has comma separated user names in one column, and email addresses:

Course 1	Bob Smith, Jane Wright	bob.smith@example.com, jane.wright@example.com
Course 2	Bob Smith	bob.smith@example.com
Course 3	Jane Wright	jane.wright@example.com

This layout is a bit weird if the lists are long and you want to associate the names and emails. We could follow the longer mailbox format, as specified in RFC 5322. This is particularly good if the email address isn’t obviously associated with or derived from the name (e.g. Bobby Apples s928777743@example.com)

Course 1	Bob Smith bob.smith@example.com, Jane Wright jane.wright@example.com
Course 2	Bob Smith bob.smith@example.com
Course 3	Jane Wright jane.wright@example.com

The most natural way to present this information is to have a cross product between courses and users, so each course will appear multiple times (once per user). This leaves a single item in each column, and allows you to easily sort by either course or user.

Course 1	Bob Smith	bob.smith@example.com
Course 1	Jane Wright	jane.wright@example.com
Course 2	Bob Smith	bob.smith@example.com
Course 3	Jane Wright	jane.wright@example.com

Notes from meetings

• When it comes to getting a full "spreadsheet" type dump, we can work around our normal 2s rule if needs be
• We are interested in "who's been using the tool since June X"
  • Therefore the critical date is not created by
  • It's activity since the period
  • Activity is logging in, or annotating
  • Our "Billable" metric is pretty poor in this situation
  • As we are mostly interested in recent activity for this, our "Active" metric is pretty much a dead on match

What does "who's been using the tool since June X" mean?

In the case of teachers, we mean teachers who are in active groups of some kind. So the group must have activity within the time period, rather than the teacher.

An example of what this might look like (US only right now) is here:

• https://report.hypothes.is/question/124-prototype-teachers-details-in-active-groups?public_id=us.lms.org.TARKQ36QQAGEfIowmGOuGQ&group_type=course&activity_since=thisweek

For students, it's not clear yet whether we want any metrics other than the annotation counts etc. For example, an active count of users in the timeframe would be difficult for us, as the timeframe is unpredictable.

[jon-betts]

We suspect there's an inefficiency in the existing process. It seems like there's some manual work being done to process one of our reports to eventually produce something a lot like the other reports we already give.

There might not be much of a point in chasing this down right now as this PRD might be about replacing this process. It's mentioned that the manual process is a pain, but it doesn't actually suggest replacing it.

If it does, this is resolved. If it's not then it's not.

[nairiboo]

@jeremydean Also with this ticket, we should be in a position close to closing, do you have any final feedback to share on it? Thanks!