CodePulse

Analytics pipeline that processes GitHub Archive data to surface insights about developer activity and repository health.

Dashboard · dbt Docs

Architecture

GitHub Archive (BigQuery public dataset)
        ↓
   dbt models
        ↓
  Looker Studio

Pulls 7 days of GitHub events (~25M rows), runs daily via GitHub Actions.

Models

Staging

• stg_github_events - Raw events from GitHub Archive
• stg_github_pushes - Push events with bots filtered out

Analytics

• developer_activity_metrics - Developer stats: push counts, streaks, activity tiers
• repository_health_scores - Repo scoring: bus factor, retention, health tiers
• contribution_patterns - Activity heatmap by hour/day
• event_type_trends - Event composition over time
• daily_activity / event_breakdown - Summary tables

Local Setup

cd codepulse_analytics
export GCP_PROJECT_ID=your-project
export GOOGLE_APPLICATION_CREDENTIALS=/path/to/key.json
dbt run --profiles-dir .

Stack

• dbt + BigQuery
• GitHub Actions (scheduled runs)
• Looker Studio (dashboards)