Skip to content

Repository files navigation

Study Habits, Digital Tools & Academic Performance

End-to-End Student Behavior Analytics Platform

This project implements a complete data engineering and analytics pipeline designed to analyze how study habits, stress levels, and digital learning tools influence student productivity and academic outcomes.

The system integrates multiple datasets, performs automated ETL processing, stores structured data in a PostgreSQL data warehouse, and produces analytical insights through Tableau dashboards.

The project demonstrates a production-style analytics architecture combining data engineering, SQL analytics, and business intelligence visualization.


Project Objectives

The primary goal of this project is to build a data-driven analytics platform capable of identifying behavioral patterns in student learning.

The platform analyzes relationships between:

  • Study hours ( related toi how many hours)
  • Study consistency
  • Stress levels
  • Digital tool usage
  • AI tool adoption
  • Academic performance indicators

The final system enables behavioral analytics and educational insights that help identify:

  • High engagement students
  • Students at academic risk
  • Impact of digital tools on productivity
  • Stress-productivity relationships
  • Study behavior trends over time

High-Level System Architecture

CSV / Survey Datasets ↓ Python ETL Pipelines ↓ PostgreSQL Data Warehouse ↓ SQL Analytical Views ↓ Tableau Dashboard

The architecture follows a modern analytics pipeline structure, similar to those used in real-world data teams.


Core Design Principles

Behavioral Analytics

The platform focuses on student behavior rather than only final grades, allowing deeper understanding of:

  • learning habits
  • productivity fluctuations
  • stress impact
  • digital tool usage

Multi-Dataset Integration

The platform integrates multiple datasets:

Dataset Purpose
Student Study Survey Weekly study behavior
Student Performance Dataset Academic performance indicators
Student Stress Dataset Psychological and lifestyle indicators

Data Engineering Pipeline

The project implements a modular ETL pipeline that performs:

  • Data ingestion
  • Data cleaning
  • Schema alignment
  • Data validation
  • Warehouse loading

Technology Stack

Layer Technology
Data Processing Python
Data Analysis Pandas
ETL Pipeline Custom Python Scripts
Database PostgreSQL
SQL Analytics PostgreSQL Views
Visualization Tableau
Version Control Git + GitHub
CI/CD GitHub Actions
Environment Management Python venv

Repository Structure

The architecture follows a modern analytics pipeline structure, similar to those used in real-world data teams.


Core Design Principles

Behavioral Analytics

The platform focuses on student behavior rather than only final grades, allowing deeper understanding of:

  • learning habits
  • productivity fluctuations
  • stress impact
  • digital tool usage

Multi-Dataset Integration

The platform integrates multiple datasets:

Dataset Purpose
Student Study Survey Weekly study behavior
Student Performance Dataset Academic performance indicators
Student Stress Dataset Psychological and lifestyle indicators

Data Engineering Pipeline

The project implements a modular ETL pipeline that performs:

  • Data ingestion
  • Data cleaning
  • Schema alignment
  • Data validation
  • Warehouse loading

Technology Stack

Layer Technology
Data Processing Python
Data Analysis Pandas
ETL Pipeline Custom Python Scripts
Database PostgreSQL
SQL Analytics PostgreSQL Views
Visualization Tableau
Version Control Git + GitHub
CI/CD GitHub Actions
Environment Management Python venv

Repository Structure

study_tools_survey │ ├── backend │ │ ├── etl │ │ baseline_etl.py │ │ weekly_etl.py │ │ student_performance_loader.py │ │ stress_loader.py │ │ utils.py │ │ │ ├── pipeline │ │ orchestrator.py │ │ │ ├── sql │ │ schema.sql │ │ analytics_views.sql │ │ refresh_views.sql │ │ │ ├── run_pipeline.py │ └── requirements.txt │ ├── data │ ├── layer2_student_performance │ │ student-mat.csv │ │ student-por.csv │ │ │ └── layer3_stress_dataset │ stress_students.csv │ ├── dashboards │ study_behavior_dashboard.twb │ ├── .github │ workflows │ pipeline.yml │ └── README.md


Data Warehouse Design

The PostgreSQL warehouse contains multiple relational tables designed for behavioral analytics.

Main Tables

Table Description
baseline_cohorts Student baseline academic context
weekly_observations Weekly behavioral survey responses
student_performance External academic performance dataset
student_stress_context Student lifestyle and stress dataset
cohort_weekly_metrics Derived analytics table

Example Schema

baseline_cohorts │ └── cohort_key │ ▼ weekly_observations

student_performance student_stress_context

The schema allows:

  • cohort-level analysis
  • weekly behavioral tracking
  • multi-dataset correlation

ETL Pipeline

The ETL pipeline is implemented in Python and consists of several modular scripts.

Baseline ETL

Processes baseline survey responses and loads them into:

Weekly ETL

Processes weekly behavioral survey responses and loads them into:


Student Performance Dataset Loader

Loads external academic performance dataset.


Student Stress Dataset Loader

Loads psychological and lifestyle dataset.


⚙ Pipeline Orchestration

To simplify execution, the entire ETL workflow can be executed using a single command.

The orchestrator performs:

Baseline ETL
Weekly ETL
Student performance dataset load
Stress dataset load
SQL analytics refresh


Analytical SQL Layer

The analytics layer generates behavioral insights through SQL views.

Engagement Index

engagement_index = productivity_level * 0.7 + (10 - stress_level) * 0.3

Academic Risk Classification

CASE WHEN stress_level >= 4 AND total_hours_this_week < 15 THEN 'HIGH_RISK'

WHEN stress_level >= 3 THEN 'MEDIUM_RISK'

ELSE 'LOW_RISK' END


Analytics Views

View Purpose
engagement_metrics Student engagement score
academic_risk_analysis Risk classification
tool_adoption_analysis Digital tool usage patterns
stress_behavior_analysis Stress behavior relationships

Tableau Dashboard

The final analytics results are visualized using Tableau.

The dashboard includes:

KPI Cards

  • Total Students
  • Average Study Hours
  • Average Productivity
  • Average Stress

Visualizations

  • Study Hours vs Productivity
  • Stress vs Productivity
  • Academic Risk Distribution
  • Digital Tool Adoption by Program
  • Weekly Engagement Trends

Interactive Filters

Users can filter results by:

  • Course program
  • University type
  • State
  • Risk category

Automation

The project supports automated execution of the ETL pipeline.

Automation features include:

  • Pipeline orchestration script
  • Automated SQL view refresh
  • Scheduled pipeline execution
  • Logging support

CI/CD Pipeline

The repository includes a GitHub Actions workflow that validates the pipeline.

The CI pipeline performs:

Install dependencies
Run ETL scripts
Validate database connections
Execute analytics SQL

This ensures the pipeline remains stable during development.


Example Pipeline Output

Baseline ETL completed. Accepted rows: 89 Rejected rows: 13

Weekly ETL completed. Accepted rows: 241 Rejected rows: 19

Loading student performance datasets... Inserted 1044 rows.

Loading stress dataset... Inserted 2000 rows.

REFRESH MATERIALIZED VIEW


Future Improvements

Potential future improvements include:

  • Real-time streaming ingestion
  • Airflow pipeline orchestration
  • Predictive modeling for academic risk
  • Machine learning models for performance prediction
  • API layer using FastAPI
  • Cloud deployment

Academic Value

This project demonstrates skills in:

  • Data Engineering
  • SQL Analytics
  • Data Warehousing
  • ETL Pipeline Design
  • Business Intelligence
  • Behavioral Data Analysis

Project Summary

This system implements a full analytics pipeline that transforms raw behavioral datasets into actionable educational insights.

The platform demonstrates how data engineering pipelines and business intelligence tools can be applied to analyze student learning behavior and productivity trends.


Author

Jayesh Kalla Rao
Computer Science and Engineering
Amrita Vishwa Vidyapeetham


Final Status

✔ Data Engineering Pipeline Complete
✔ PostgreSQL Warehouse Implemented
✔ Analytical SQL Layer Complete
✔ Tableau Dashboard Functional
✔ Automation Implemented
✔ CI/CD Pipeline Integrated

Project ready for demonstration and portfolio use.


🔗 Deployed Dashboard link

👉 View on Tableau Public


Project Status

Build Status Python License

Team

  • JAY-cloudbuster (Lead Developer)
  • Kanishthika11 (Data Engineer)
  • Mayukasri06 (Documentation & Data)
  • Srijak1202 (Testing & QA)
  • suhas107 (DevOps & CI/CD)

About

No description, website, or topics provided.

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages