Learner Performance Analytics

A data science project that investegates learner performance and engagement in courses using exploratory data analysis and Neural Networks

Dashboard Preview:

Model Preview:

The proposed neural network was trained to simultaneously perform two related tasks:

• Classification: Predict the learner's performance category (Excellent, Good, Average, Needs Support).
• Regression: Predict the learner's final performance score.

Dataset overview:

Number of records: 5010 (5000 after cleaning)

Number of columns: 18 Columns

Data question: How can we increase the engagement and learner performance in courses

Technologies used:

• Python
• Pandas (Data cleaning and descriptive analysis)
• Numpy (Converting pytorch tensors to numpy array for scikit-learn)
• Matplotlib (Visuallizing data)
• Seaborn (Visuallizing data)
• Scikit-learn (Data preprocessing pipelines and evaluation metrics)
• Pytorch (Model development and training)
• Torchview (Visualling model architecture)

Data preprocessing:

• Filling missing values in the data with mean for numerical columns and mode for categorical columns
• Removing duplicates
• Converting videos column from categorical (11 videos) to numerical (11)

Exploratory Data Analysis:

• Distribution of final scores
• Probabillity of courses completion rate
• Correlation analysis using heatmaps
• Engagement analysis

Model preview:

A NN model using pytorch was trained with two prediction heads (A classification head and regression head) and 3 shared layers

What the model predicts:

• Engagement score
• Final score
• Learner Performance

Evaluation metrics

Classification evaluation metrics:

• Precision
• Accuracy
• Recall

Regression evaluation metrics:

• MSE
• MAE
• R2 score

Model's evaluation

Note: The model achieved balanced performance across the four learner categories, with F1-scores ranging from 0.63 to 0.75. Performance differences among classes are primarily influenced by class imbalance and the complexity of distinguishing students with similar learning characteristics.

Installation

1. Clone the repository

git clone https://github.com/youssef124wael/Learner-Performance-Analytics
cd learner-performance-analytics

2. Create a virtual environment (recommended)

python -m venv venv

Activate the environment:

Windows

venv\Scripts\activate

Linux/macOS

source venv/bin/activate

Git bash

source venv/Scripts/activate

3. Install the required dependencies

pip install -r requirements.txt