This project is about understanding how students drop out from online courses.
In many online platforms, a lot of students join courses but many do not complete them.
The goal is not prediction or advanced AI, but simple statistics and visualization.
- Divide students into early, mid, and late dropouts
- Compare engagement levels between groups
- Analyze relationship between activity and dropout timing
A synthetic dataset was used due to privacy and availability constraints.
The dataset includes:
- Course duration
- Last active day
- Number of assignments submitted
- Total video watch time
- Average quiz score
Google Colab, NumPy, Pandas, Matplotlib, Seaborn, Gradio
- Bar plots for dropout count
- Box plots for engagement comparison
- Violin plots for distribution
- Scatter plots for engagement vs dropout
- Correlation heatmap
- Many students drop out early
- Higher engagement correlates with course completion
- Quiz score alone is not a strong indicator
- Engagement matters more than performance
A simple interactive UI allows:
- Selecting dropout phase
- Choosing plot types
- Viewing sample data
Shubh Raj
Content here
Content here
Content here
Content here
Content here
Content here