1. Dataset Name

tm_learnings_v2

2. Description

An upskilling-activity-level dataset that combines three learning and development sources into a single unified table: course completions from the LMS, project-based upskilling from team membership records, and development plan completions. Each row represents one upskilling activity, including the source type, completion date, associated skill, and full employee context. This dataset powers course completion counts, unique participant counts, skills gained, and monthly trend metrics on the TM Skills dashboard.

3. Source Tables

• customer_views.course_attendance (Course branch — LMS course completion records, UNNEST on courses.skills_json for skill extraction, LEFT JOIN to customer_views.courses for course metadata)
• customer_views.project_teammembers_details (Project branch — project team member completion records, WHERE project_completed_ts IS NOT NULL)
• customer_views.profile_upskilling (Upskilling branch — development plan completions, UNNEST on calibration_json for skill extraction)
• customer_views.employee (aliased c — joined to all three branches on profile_id to enrich with employee context; joined twice in the outer CTE structure)

4. Grain

One row per upskilling activity per employee per skill per source type. Because skills are unnested from JSON arrays in the Course and Upskilling branches, a single course with three associated skills will produce three rows. The source_type column ('Course', 'Project', 'Upskilling') distinguishes which UNION ALL branch each row originates from.

5. Dashboards Supported

• TM Skills Dashboard

6. Metrics Supported

7. Known Exclusions / Filters Applied

• Course branch: WHERE completion_date IS NOT NULL in course_attendance. Skills are UNNEST'd from courses.skills_json — rows without matching course skills will not appear. completion_percentage is hardcoded to 1 for all course rows (fully completed by definition of having a completion_date).
• Project branch: WHERE project_completed_ts IS NOT NULL in project_teammembers_details. completion_percentage is hardcoded to 1. project_completed_ts is used as the completion_date for this branch.
• Upskilling branch: no explicit completion filter in the UNION ALL — filtering to fully completed activities (completion_percentage = 1.0) is applied at the widget level via implicit_filters. Skills are UNNEST'd from calibration_json.
• source_type is a derived column: 'Course' (course_attendance branch), 'Project' (project_teammembers_details branch), 'Upskilling' (profile_upskilling branch). Available as a drilldown dimension for combined completion metrics.
• The employee context (business_unit, lob, location_country, seniority, etc.) is joined from employee via profile_id. Project-branch rows carry both employee-level and project-team-member-level attributes (prefixed project_teammember_*).
• Employees who Completed Development Plan - All Time and its Business Unit breakdown have no date range filter applied — they are all-time snapshots. All other metrics filter by completion_date (Course/Upskilling) or project_end_ts (Project).

8. Schema