Exploration of different electricity demand forecasting methods for PSEG
This mini-project was undertaken as part of Forecasting Methods at HEC Montreal.
The goal of this project was to investigate various forecasting methods for predicting the electricity demand one day ahead for the PSEG region (near New York and Philadelphia) in the USA.
The following scripts were created:
E01_load_demand.R: to extract the electricity demand time series as well as relevant weather data
E01_dshw_tbats.R: exponential smoothing methods (double seasonal holt-winters and TBATS)
E01_reg_arima.R: multiple linear regression models with arima error components
E01-arima.R : ARIMA & SARIMA methods
I'd like to also give credits to Nicolas Gervais (https://github.com/nicolas-gervais) who has been another important contributor to the R code.