EdNet SaaS Learning Analytics Platform

End-to-End Analytics Engineering & BI Pipeline

📌 Business Case: Driving Retention via Engagement Modeling

• Problem: High early-stage churn (Weeks 1-4) in the tutoring SaaS platform leads to diminished Customer Lifetime Value (LTV).
• Goal: Build a scalable 3-tier (RAW-CORE-MART) data warehouse to monitor weekly retention and identify "at-risk" students based on engagement density.
• Output: A production-style Operational Command Center that generates a "weekly risk list" for targeted re-engagement campaigns.

🛠️ Tech Stack

• Data Warehousing: Snowflake (SQL) - 3-Tier Architecture
• Statistical Modeling: R (OLS Regression) - Baseline Behavioral Analysis
• Visualization: Tableau Public - Operational Dashboards

🏗️ Data Architecture & Engineering (Snowflake)

The pipeline implements SaaS-standard data governance to ensure reliability and performance:

• 01_Setup (Ingestion): Environment initialization and raw data auditing with Data Quality (DQ) guards (null-rate and duplicate detection).
• 02_Core (Normalization): Event-level FACT_SOLVES with incremental processing logic. Data grain is set at the user-event-timestamp level.
• 03_Mart (Metric Layer): Aggregated tables optimized for BI performance, including Weekly Retention Cohorts and Student Feature Marts.

Engineering Highlights:

• Cost Optimization: Configured X-Small warehouse with auto-suspend and utilized Query Pruning on event dates to minimize compute credits.
• Reliability: Integrated row-count drift monitoring and freshness checks within the SQL workflow.

📈 Insights & Operationalization (R)

Instead of a static study, the R model serves as an Explainable Baseline for business intervention:

• Core Insight: Quantified that Active Days per Week is the primary driver of student productivity (Total Questions Solved).
• Actionability: The model identifies "low-frequency" user segments, pushing Risk_Flags back to the Snowflake Mart for automated CRM triggers.

📊 Business Intelligence (Tableau)

The dashboard provides a Macro-to-Micro drill-down for operations teams:

• Weekly Retention Heatmap: Tracks 35-week user decay to pinpoint critical drop-off windows.
• Engagement Scatter: Linked visuals that allow operators to filter engagement behavior by acquisition cohort.
• 🔗 View Live Operational Dashboard

📂 Repository Structure

├── sql/                # Production SQL: 01_Setup, 02_Core, 03_Mart
├── r_analysis/         # R scripts for OLS baseline & risk segmentation
├── tableau/            # Dashboard documentation & screenshots
└── data_samples/       # Schema definitions & top-10 row samples for reference