Topic 3 – GDP Time Series for Multiple Countries (Kaggle)

Dataset

• Source: GDP Timeseries Data for various Countries – Kaggle
• Link: https://www.kaggle.com/datasets/iamtushara/gdp-timeseries-data-for-various-countries

Download Instructions

1. Open the dataset link above.
2. Click "Download".
3. Extract to data/.
4. Use the main GDP CSV (e.g. gdp_timeseries.csv).

Data Loading

import pandas as pd
df = pd.read_csv("data/gdp_timeseries.csv")  # adjust filename
country = df[df["Country"] == "Vietnam"].copy()  # example filter
country["Year"] = pd.to_datetime(country["Year"], format="%Y")
country = country.set_index("Year").sort_index()

Implementation Steps

1. Data Exploration

• Load dataset and explore structure
• Select 1-3 countries for analysis (e.g., Vietnam, USA, China)
• Inspect GDP values and time range
• Check for missing values and data quality

2. Exploratory Data Analysis (EDA)

• Plot GDP time series for each country
• Compare trends across countries (side-by-side plots)
• Calculate growth rates (year-over-year, CAGR)
• Identify structural breaks or regime changes
• Perform time series decomposition

3. Stationarity Analysis

• Test each country's GDP series for stationarity (ADF test)
• Apply differencing if needed (GDP is typically non-stationary)
• Consider log transformation for variance stabilization
• Check for unit roots

4. Model Building

• Univariate Models (per country):
• ARIMA models
• Consider trend models if strong deterministic trend
• Comparative Analysis:
• Compare model performance across countries
• Identify common patterns or differences
• Advanced (optional):
• VAR models for multivariate analysis
• Cointegration tests if analyzing multiple countries together

5. Model Evaluation

• Split data temporally (e.g., last 5-10 years as test set)
• Generate forecasts for each country
• Calculate metrics: MAE, RMSE, MAPE
• Compare forecast accuracy across countries
• Visualize forecasts with confidence intervals

6. Comparative Analysis

• Compare GDP growth trends
• Analyze forecast accuracy differences
• Discuss economic factors affecting each country
• Generate comparative forecast plots

Expected Deliverables

1. EDA Report:

• GDP plots for each country
• Growth rate analysis
• Comparative visualizations
• Stationarity test results

1. Model Results:

• Model parameters for each country
• Performance metrics comparison table
• Forecast plots (individual and comparative)
• Discussion of differences

1. Code:

• Python notebook with all analyses
• Functions for country comparison
• Visualization utilities

Tips

• GDP data is typically annual, so fewer data points - be careful with model complexity
• GDP usually has strong upward trend (non-stationary)
• Consider log transformation to model growth rates
• Compare countries with similar economic characteristics for meaningful insights
• Use growth rates (first difference of log GDP) for stationary analysis
• External factors (recessions, policy changes) may cause structural breaks