This project analyzes daily vehicle traffic data from the Baregg Tunnel (2003–2005) to develop and evaluate forecasting models. A Naïve benchmark model is compared against a Linear Regression model incorporating trend and weekly seasonality. The dataset exhibits strong weekly seasonality and a mild upward trend. Using a validation period (July 2005 – November 2005), model performance is evaluated through RMSE, MAE, MAPE, and MASE. Results show that the Linear Regression model significantly outperforms the Naïve approach by effectively capturing weekly traffic patterns and underlying trend components. Residual diagnostics confirm that model assumptions are satisfied, indicating reliable forecasts. This analysis demonstrates how incorporating seasonality and trend improves forecasting accuracy in real-world transportation data.

Primera tarea de Programación

simp59

Demonstration plots

Fundamentos de programación