Student Academic Performance in Ethiopia: Prediction, Key Drivers, and Statistical Insights
# Predicting Ethiopian Student Academic Performance Using Machine Learning
---
## ๐ Project Overview
Education shapes the future of individuals and nations, yet student performance is often discussed emotionally rather than scientifically.
As someone who personally experienced the pressure of the Ethiopian Grade 12 national entrance examination, I wanted to investigate a deeper question:
> **What truly drives student success?**
> Is it family background? School resources? Student behaviour? Or something else entirely?
To answer this, I analyzed a large Ethiopian student dataset using **Machine Learning, Statistical Testing, and Predictive Modeling**.
This project combines data science with a real educational challenge affecting thousands of students.
---
## ๐ฏ Objectives
This project was designed to answer five major questions:
- Can student academic performance be predicted accurately?
- Which factors most strongly influence performance?
- Do school resources matter more than family background?
- Does student behaviour improve prediction accuracy?
- Are there statistically significant differences across gender, region, school type, and internet access?
---
## ๐ Dataset
**Source:** Kaggle โ Ethiopian Students Dataset
The dataset employed here is a simulated version of a real Ethiopian education dataset, specifically crafted to reflect the characteristics and challenges of actual student records. Real government data of this nature is confidential and inaccessible, making this simulated dataset an invaluable resource for this study.The raw dataset is a vary rich data covering demographics, parental background, school characteristics, attendance, homework, and national examination results.
Due to GitHub file size limitations, the dataset is hosted externally.
Processed dataset used in this project is loaded directly within the notebook.
Original source: [
kaggle.com]
### Raw Dataset Size โฆ