A Multidimensional Integrated Dataset of Behavioral and Performance for Academic Student (Education)

Description

This dataset comprises 19,222 observations synthesized through a robust Medallion Architecture in a SQL Server environment. By integrating data from three primary educational sources Student Productivity, Exam Prediction, and Student Performance it provides a multidimensional view of academic success factors. Key Analytical Features: Behavioral: Study hours, sleep patterns, and digital consumption (gaming/social media). Physiological: Self-reported stress levels and focus metrics. Academic: Historical performance (Previous_Grades) and attendance rates. Demographics: Age-binned cohorts and gender distribution. The dataset serves as a high-integrity foundation for supervised machine learning, offering both binary (is_passing) and multi-class (final_grade) target variables to support nuanced classification tasks.

Steps to reproduce

The transformation of raw data into the final analytical "Gold Layer" followed a rigorous three-stage engineering process: Bronze (Ingestion): Raw data from the three distinct educational sources was ingested into SQL Server without schema modifications. This stage preserved the original fidelity of the digital distraction records, performance logs, and exam metadata. Silver (Refining & Integration): * Normalization: Variables across disparate sources were standardized into consistent formats (e.g., converting study time to float). Data Integrity: Imputation of missing values and removal of outliers. Join Logic: Records were unified using unique identifiers to create a cohesive student profile, ensuring behavioral data correctly mapped to academic outcomes. Gold (Optimization): Feature Engineering: Age was binned into five strategic cohorts for better categorical analysis. Target Labeling: Created the is_passing binary flag based on specific grade thresholds. Anonymization: Removed PII (Personally Identifiable Information) to ensure the dataset meets ethical research standards for sharing and modeling.

Institutions

• Mansoura University

  Dakahlia, Al Mansurah
• New Mansoura University

  Dakahlia, Al Mansurah

Categories

Related Links

Licence