Statistical Analysis & Predictive Modeling of Youth Unemployment in Morocco (2000-2024)
# Statistical Analysis & Predictive Modeling of Youth Unemployment in Morocco (2000-2024)
**End-of-Module Project: Statistical Inference & Data Analysis**
*École Mohammadia d'Ingénieurs (EMI)*
---
## 📖 Project Overview
This project applies **Data Science** and **Statistical Inference** techniques to analyze the structural determinants of youth unemployment in Morocco. Using a dataset spanning 24 years (2000-2024), we modeled the complex interactions between macro-economic variables and labor market dynamics, comparing the results with Sub-Saharan Africa trajectories.
---
## 👨✈️ My Contribution: Team Lead & Project Pilot
As the **Group Leader**, I was responsible for the project architecture, workflow coordination, and ensuring the scientific rigor of our approach.
My specific technical contribution focused on **Part 2: Macro-economic Modeling & Analysis**. I led the investigation into how major economic indicators drive—or fail to drive—youth employment.
### 🧠 AI & Statistical Concepts Applied
We moved beyond descriptive statistics to apply **Machine Learning** and **Inferential** concepts:
* **📉 Linear Regression Models (OLS):** Quantifying the elasticity of unemployment relative to economic shocks.
* **🔮 Predictive Analysis:** Forecasting trends based on historical volatility and "monotony" analysis.
* **✅ Hypothesis Testing:** Validating assumptions (e.g., Phillips Curve, Okun's Law) using rigorous statistical tests (**Student, Shapiro-Wilk, Wilcoxon**).
* **📊 Correlation Matrix Analysis:** Identifying latent relationships between 8 key variables.
---
## 🔍 Key Findings: Macro-economic Analysis
My analysis highlighted a fundamental **"Moroccan Paradox"**: High investment rates but low job creation for youth.
### 1. Agriculture: The Stochastic Regulator 🌾
> *"The economy breathes to the rhythm of the rainfall."*
* **Distribution:** **Bimodal distribution**, reflecting the binary nature of the economy (drought vs. abundance).
* **Regression R …