KinushDK/YouTube-Kenya-Analytics-Pipeline

# 📺 YouTube Kenya Analytics Pipeline A production-grade data engineering pipeline that tracks, analyses, and visualises performance metrics for 9 major Kenyan YouTube channels — built entirely on Google Cloud Platform with a modern ELT stack. > **Live Dashboard →** YouTube Kenya Analytics Dashboard > **GitHub →** KinushDK/YouTube-Kenya-Analytics-Pipeline --- ## The Problem This Solves Content creators, media houses, and digital marketers in Kenya make publishing decisions based on gut feeling. Which day should Citizen TV upload? Does Sunday afternoon outperform Monday morning? Does a 3-minute video outperform a 15-minute one? Is engagement rate a better signal than raw views? This pipeline answers those questions with data — pulling daily snapshots from the YouTube API, transforming them through a structured data model, and surfacing the insights in a live dashboard that updates automatically every day. --- ## Real Business Insights From This Data Running this pipeline for 90 days revealed patterns that matter to any Kenyan media operation: **Sunday afternoon is the best time to post** — the data shows Sunday outperforms every other day by a significant margin, with afternoon slots (12pm-6pm) generating the highest average views across all channels. **Music has a higher engagement rate than News** — despite News & Politics having 12.3 million total views, Music content achieves a 2.47% engagement rate vs 0.95% for news. Volume and engagement tell different stories. **63% of videos are Brief (1-5 minutes)** — the shift toward short-form content is clear across all channels regardless of niche. **Citizen TV Kenya dominates** at 6.65 million subscribers and 3 billion total views — the largest Kenyan YouTube channel by every metric. --- ## Architecture ``` YouTube Data API v3 │ │ (daily at 06:00 EAT) ▼ Python Ingestion Script │ ├──► Google Cloud Storage (raw JSON, partitioned by date) │ └──► BigQuery raw dataset │ │ (dbt transforms) ▼ BigQuery staging …