🌍 Global GDP Per Capita Analysis & Forecast (2000–2030)

A data science project analyzing global economic performance using the Penn World Table dataset.
The project explores GDP per capita trends, country clustering, statistical testing, and future predictions using machine learning.

---

📊 Project Goals

This project aims to:

• Analyze global GDP per capita trends (2000–2018)
• Identify economic groupings of countries using clustering
• Compare developing vs developed economies statistically
• Predict future GDP per capita trends using regression
• Visualize global economic patterns

---

📁 Dataset

Source: Penn World Table (PWT 10.01)

The dataset contains macroeconomic indicators for countries worldwide.

Key variables used:

Variable	Description
country	Country name
year	Year of observation
rgdpo	Real GDP (output-side)
pop	Population
gdp_per_capita	Calculated as rgdpo / pop

The analysis focuses on the 2000–2018 time range.

---

🧹 Data Preparation

Data preprocessing steps included:

• Filtering data between 2000–2018
• Removing missing values
• Removing rows with population = 0
• Creating a new feature:

gdp_per_capita = rgdpo / pop

This metric allows comparison of economic productivity per person across countries.

---

📈 Exploratory Data Analysis

Descriptive statistics were used to understand global economic patterns.

Metrics calculated:

• Mean GDP per capita
• Median GDP per capita
• Mode GDP per capita
• Quartiles (25%, 50%, 75%)

These provide insights into global standards of living and economic inequality.

---

🤖 Machine Learning: Country Clustering

To identify economic groups, K-Means clustering was applied.

Steps:

1. Calculate average GDP per capita per country
2. Standardize values using StandardScaler
3. Apply K-Means clustering (k = 4)

Clusters represent different economic categories such as:

• Low-income economies
• Emerging economies
• Upper-middle income countries
• High-income economies

Visualization example:

• Scatter plot of GDP per capita vs population
• Countries grouped by cluster

---

📊 Statistical Testing

The project tests whether GDP per capita distributions differ between economic groups.

Normality Test

Shapiro–Wilk Test

Used to check whether the GDP distribution is normal.

Result:

• GDP per capita data is not normally distributed.

Group Comparison

Mann–Whitney U Test

Used instead of a t-test because the data is non-normal.

This test compares GDP distributions between developing and developed country clusters.

---

📉 Global GDP Growth Over Time

Year-over-year GDP growth was calculated using:

GDP Growth = (GDP_t − GDP_(t−1)) / GDP_(t−1)

The average global GDP growth was then visualized over time to observe long-term economic trends.

---

🌍 Top Performing Countries

Countries were ranked by average GDP per capita.

The project identifies the top 10 performing economies based on average GDP per capita between 2000–2018.

Example country comparison:

• United States
• Germany
• Egypt
• India

---

📈 GDP Forecasting (Machine Learning)

A Linear Regression model was used to predict GDP per capita trends.

Example prediction:

Finland GDP per Capita (2000–2030)

Model workflow:

1. Train model using historical data (2000–2018)
2. Use year as predictor
3. Predict GDP per capita through 2030
4. Compare actual vs predicted values

This demonstrates a basic economic forecasting approach.

---

📊 Example Visualizations

The project includes:

• GDP clustering scatter plots
• Global GDP growth over time
• Country GDP comparisons
• GDP prediction graphs

Libraries used:

• Matplotlib
• Seaborn

---

🧰 Technologies Used

Technology	Purpose
Python	Programming language
Pandas	Data manipulation
NumPy	Numerical computing
Matplotlib	Data visualization
Seaborn	Statistical visualization
Scikit-learn	Machine learning
SciPy	Statistical testing

---

▶️ Running the Project

Clone the repository

git clone https://github.com/yourusername/gdp-analysis-project.git
cd gdp-analysis-project

Install dependencies

pip install pandas numpy matplotlib seaborn scikit-learn scipy

Run the analysis

python gdp_analysis.py

Make sure the dataset file pwt1001.csv is located in the project directory.

---

📌 Key Insights

• GDP per capita varies widely between countries.
• Machine learning clustering reveals clear economic groupings.
• GDP distributions are non-normal, requiring non-parametric tests.
• Linear regression can approximate long-term economic trends, though it cannot capture complex economic shocks.

---

🚀 Future Improvements

Potential extensions:

• Add more economic indicators (inflation, unemployment, education)
• Use time series models (ARIMA, Prophet, LSTM)
• Build an interactive dashboard (Plotly / Streamlit)
• Apply multi-feature clustering for deeper economic classification

---

👩‍💻 Author

Olga Chitembo

This project is part of my data science and analytics portfolio, demonstrating skills in:

• Data analysis
• Statistical testing
• Machine learning
• Economic data interpretation
• Data visualization