Econometrics & Forecasting

Econometrics
Forecasting
Time Series
Applied forecasting projects using ARIMA, VAR, and time series methods.

A collection of applied forecasting projects using economic time series, model diagnostics, and forecast evaluation. These projects focus on understanding economic behavior over time and comparing models based on interpretability and out-of-sample performance.

Projects

Forecasting Monthly U.S. Retail Sales

R · Time Series · Decomposition · ARIMA · Forecasting

Modeled monthly U.S. retail sales to study long-run trends, seasonality, and short-term forecast behavior.

Highlights

  • Tested stationarity using ADF tests
  • Compared linear and quadratic trend models
  • Evaluated additive vs. multiplicative decomposition
  • Built 12-month retail sales forecasts

GitHub ↗

U.S. Labor Market Forecasting

R · ARIMA · VAR · Granger Causality · Forecast Evaluation

Analyzed jobless claims and temporary help employment to evaluate whether labor market indicators improve short-term forecasting.

Highlights

  • Built structured ARIMA models
  • Compared manual and automatic model selection
  • Estimated VAR models
  • Used impulse response functions and Granger causality tests

GitHub ↗

Forecasting the USD/COP Exchange Rate

R · ARIMA · ETS · NNETAR · Prophet · Forecast Combination

Compared forecasting models for the USD/COP exchange rate using out-of-sample forecast accuracy.

Highlights

  • Modeled monthly exchange rate dynamics
  • Compared ARIMA, ETS, Holt-Winters, NNETAR, and Prophet
  • Evaluated RMSE, MAE, and MAPE
  • Built a forecast combination model

GitHub ↗

Skills Demonstrated

  • R
  • fpp3
  • forecast
  • quantmod
  • ARIMA
  • VAR
  • ETS
  • Prophet
  • Granger Causality
  • Forecast Evaluation
  • Time Series
  • ← Return to all projects