This is a prediction to predict youth unemployment in Africa. 5t is important to note that it is built on synthetic data, and as such should not be assumed as the real situation on ground.
README_YouthUnemployment.md
# Youth Unemployment Prediction in Africa (Synthetic DataSet)
## Project Overview
This project predicts the **youth unemployment rate** across African countries using a synthetic dataset of 40 country-level observations, with the goal of helping policymakers and development organizations understand which macroeconomic factors are most associated with youth joblessness. The dataset includes GDP per capita, an education index, and urban population percentage as predictors. A Linear Regression model was built to estimate youth unemployment rate, achieving an **R² of 0.72** (explaining 72% of the variance in unemployment) with a **mean absolute error of about 2.1 percentage points** on held-out test data.
## Business Understanding
The stakeholders for this project are government policymakers, regional economic bodies (e.g., the African Union, national ministries of labour), and development/NGO organizations focused on youth employment programs. Youth unemployment is a major socioeconomic concern across Africa, contributing to lost economic output, increased migration pressure, and social instability, making it a key priority for policy intervention. Understanding which macroeconomic levers (e.g., GDP growth, education investment, urbanization) are most strongly associated with unemployment outcomes helps stakeholders prioritize where to direct limited public investment and design targeted interventions rather than broad, unfocused programs.
## Data Understanding
The dataset (`youth_unemployment_africa.csv`) is **synthetic** and contains **40 country-level records** (one row per country, all for the year 2021), with columns for `Country`, `Year`, `GDP_Per_Capita`, `Education_Index`, `Urban_Population_Percent`, and the target variable `Youth_Unemployment_Rate`. Because the data is synthetic and limited to a single year, this is an important limitation: the model captures patterns designed into the synthetic generation process rather than …