- Python 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| .idea | ||
| y | ||
| main.py | ||
| README.md | ||
SynEdu-HEDL: Synthetic Higher Education Digital Learning Dataset
📊 Dataset Overview
SynEdu-HEDL is a comprehensive synthetic dataset designed to simulate realistic learner behavior, academic performance, and digital engagement patterns in higher education. This dataset enables privacy-preserving research in learning analytics, educational data mining, and decision support systems without compromising real student data.
Key Features
- Privacy-Preserving: 100% synthetic data with no real student information
- Realistic Correlations: Patterns mirror actual educational environments
- Multi-Table Structure: 6 interconnected tables with referential integrity
- Scalable Design: Configurable to generate datasets of various sizes
- Research-Ready: Includes derived metrics and outcome labels for ML/Analytics
🎯 Purpose
This dataset supports research in:
- Learning Analytics and Educational Data Mining
- Predictive Modeling of Student Performance
- Digital Engagement Pattern Analysis
- Educational Intervention Effectiveness
- Privacy-Preserving Educational Research
- Algorithm Development for Learning Management Systems
📈 Dataset Scale
| Component | Count | Description |
|---|---|---|
| Students | 20,000 | Synthetic learners with demographic profiles |
| Courses | 180 | Various course types and difficulty levels |
| Faculty | 420 | Teaching staff (implied, not explicitly modeled) |
| Semesters | 6 | Academic periods spanning 3 years |
| LMS Events | 120,000+ | Digital interaction records |
| Assessments | ~300,000 | Performance records across courses |
🏫 Institutions Modeled
- Engineering universities
- Multidisciplinary higher education institutions
- Similar to Indian private/state universities
📁 Dataset Structure
1️⃣ STUDENT_PROFILE Table
Contains demographic and background information for synthetic learners.
| Attribute | Type | Description | Values/Range |
|---|---|---|---|
| student_id | Integer | Unique synthetic learner ID | 1-20,000 |
| gender | Categorical | Gender identity | Male/Female/Other |
| age | Integer | Student age | 18-26 |
| admission_type | Categorical | Admission category | Merit/Management/Transfer |
| program | Categorical | Degree program | B.Tech/M.Tech/MCA |
| specialization | Categorical | Academic specialization | AI, DS, CSE, IT |
| socioeconomic_index | Integer | Proxy for access/resources | 1-5 (Low-High) |
| first_gen_learner | Boolean | First-generation student | True/False |
| baseline_digital_literacy | Float | Initial tech readiness | 0-1 |
| enrollment_year | Integer | Admission year | 2019-2022 |
2️⃣ COURSE_METADATA Table
Contains course characteristics and metadata.
| Attribute | Type | Description | Values/Range |
|---|---|---|---|
| course_id | Integer | Unique course identifier | 1-180 |
| course_type | Categorical | Course format | Theory/Lab/Project |
| credit_value | Integer | Academic credits | 2-5 |
| delivery_mode | Categorical | Instructional delivery | Blended/Online/Offline |
| assessment_weightage | JSON | Internal/External split | {"Internal": X, "External": Y} |
| difficulty_level | Integer | Academic rigor | 1-5 (Easy-Hard) |
| tech_dependency | Integer | LMS reliance | 1-5 (Low-High) |
3️⃣ LMS_INTERACTION_LOG Table
Detailed Learning Management System interaction records.
| Attribute | Type | Description | Values/Range |
|---|---|---|---|
| event_id | Integer | Unique log entry | 1-120,000+ |
| student_id | Integer | Foreign key → Student | 1-20,000 |
| course_id | Integer | Foreign key → Course | 1-180 |
| event_type | Categorical | Interaction type | Login, VideoView, Quiz, Forum, etc. |
| session_duration | Float | Time spent (minutes) | 1-120 |
| resource_type | Categorical | Content type | Video, PDF, Assignment, etc. |
| timestamp | Datetime | Interaction time | 2023-2025 |
| device_type | Categorical | Access device | Mobile/Laptop/Tablet/Desktop |
| network_quality | Integer | Connectivity proxy | 1-5 (Poor-Excellent) |
4️⃣ ASSESSMENT_PERFORMANCE Table
Student assessment scores and submission details.
| Attribute | Type | Description | Values/Range |
|---|---|---|---|
| assessment_id | Integer | Unique assessment | 1-300,000+ |
| student_id | Integer | Foreign key | 1-20,000 |
| course_id | Integer | Foreign key | 1-180 |
| assessment_type | Categorical | Assessment category | Quiz/Mid/End/Assignment |
| max_marks | Integer | Total marks | Varies by type |
| obtained_marks | Float | Student score | 0-max_marks |
| submission_mode | Categorical | Submission method | LMS/Offline |
| plagiarism_flag | Boolean | Academic integrity | True/False |
5️⃣ ENGAGEMENT_METRICS Table
Derived engagement indicators from interaction logs.
| Metric | Type | Description | Range |
|---|---|---|---|
| weekly_login_frequency | Float | LMS usage pattern | 0-20 |
| average_session_time | Float | Engagement depth | 0-120 min |
| content_completion_rate | Float | Learning persistence | 0-1 |
| forum_participation_score | Float | Collaborative learning | 0-1 |
| assessment_timeliness | Float | Self-regulation | 0-1 |
6️⃣ OUTCOME_LABELS Table
Learning outcomes and predictive labels for analytics.
| Attribute | Type | Description | Values/Range |
|---|---|---|---|
| final_grade | Categorical | Course final grade | A/B/C/D/F |
| learning_gain_index | Float | Pre-Post improvement | 0-1 |
| dropout_risk | Binary | At-risk identification | True/False |
| course_satisfaction | Integer | Student perception | 1-5 (Low-High) |
🔧 Synthetic Data Generation Logic
The dataset was generated using a hybrid rule-based and probabilistic modeling framework:
Key Generation Principles:
1. Student engagement patterns are conditioned on:
- Digital literacy levels
- Course difficulty
- Delivery mode (Blended/Online/Offline)
2. Assessment outcomes are probabilistically influenced by:
- Engagement metrics
- Attendance regularity (derived from LMS logs)
- Prior academic performance
3. Realistic correlations observed in higher education:
- Socioeconomic status ↔ Resource access
- Digital literacy ↔ LMS engagement
- Engagement metrics ↔ Academic performance
- Course difficulty ↔ Dropout risk