Instructor Effectiveness Modeling

Project Overview

This project analyzes instructor performance on an EdTech platform by evaluating data from 2,000 batches. The goal is to quantify "effectiveness" using a multi-factor composite score and use Machine Learning to categorize instructors into performance tiers (Low, Medium, High).

Dataset Features

The model processes the following student engagement and outcome metrics:

• Completion Rate: Percentage of students finishing the course.
• Dropout Rate: Percentage of students leaving the course early.
• Score Improvement: Average increase in student performance scores.
• Engagement: Average watch time, quiz scores, and assignment submission rates.
• Feedback: Average feedback scores and instructor response rates.

Methodology

1. Data Preprocessing & Scaling

To ensure all features contribute equally, variables like avg_score_improvement and avg_quiz_score were normalized to a common range.

2. Defining the "Effectiveness Score"

A custom weighted formula was developed to quantify performance:

• 30%: Completion Rate
• 20%: Retention (1 - Dropout Rate)
• 10% each: Score Improvement, Quiz Scores, Watch Time, and Assignment Submissions
• 5% each: Forum Activity and Feedback Scores

3. Machine Learning Model

• Algorithm: Random Forest Classifier.
• Logic: Instructors were grouped by their mean scores and split into 3 tiers using quantiles.
• Split: 80% Training / 20% Testing.

Results

The model achieved an Accuracy of ~92%.

Precision 0.94 Recall 0.92 F1-Score 0.92

Key Feature Importances:

1. Completion Rate (~33.9%): The strongest predictor of success.
2. Dropout Rate (~28.2%): A critical indicator of low engagement.
3. Avg Score Improvement (~10.1%): Reflects direct learning outcomes.

Critical Analysis

• Misleading Variables: Assignment submission rates might be influenced more by course difficulty than instructor skill.
• Real-world Risks: The model does not capture qualitative data like teaching style or communication skills.
• Recommendation: This should be used as a supportive tool rather than the sole metric for performance evaluation.

How to Use

1. Clone this repository.
2. Ensure you have pandas, seaborn, matplotlib, and scikit-learn installed.
3. Update the CSV file path in the first code cell of main.ipynb.
4. Run the notebook cells sequentially.