The data used for analysis that is sourcedfrom ENTSO-E, launched in 2015. It’s a Central collection of Electricity generation, transportation, and consumption data for the Pan-European market (ENTSOE, 2015). The provided dataset encompass the period from 2015 to 2024, with observations collected from over 30 countries across Europe. According to the aim of this project, only data related to Germany is separated for analysis. A total of 14966931 observations were selected related to Germany and 13 different covariates are available. It’s a time-series dataset where every column shows hourly values. The Projects and Reports are covering the following

1. Developed and compared models to forecast hourly electricity load and prices using over nine years of real-world German market data, spanning linear methods (AR, OLS) and machine learning algorithms (Deep Neural Network, Decision Trees).

2. Applied feature engineering, time series decomposition, and rolling-window estimation for dynamic model updates, improving predictive accuracy.

3. Evaluated model performance using out-of-sample MSFE and Diebold-Mariano tests, and interpreted machine learning predictions through variable importance analysis to identify key market drivers.