To build the best possible model for predicting UK electricity demand each half-hour of the day, 24 hours in advance.

Need Masters in Statistics with Data Sciencerequires coding in RStudio and a report, “To build the best possible model for predicting UK electricity demand each half-hour of the day, 24 hours in advance. You should probably use the last year of data for validation, and the earlier data for training the model. Generalized additive models (e.g. R package mgcv) and random forests (R package randomForest) are two worth considering (possibly in combination).” I have the data file in R format. So basically the topic is to build the best predictive GAM model, in order to predict electricity I need the project in a specific template we were given by the university (it is in R format). "To build the best possible model for predicting UK electricity demand each half hour of the day, 24 hours in advance. You should probably use the last year of data for validation, and the earlier data for training the model. Generalized additive models (e.g. R package mgcv) and random forests (R package randomForest) are two worth considering (possibly in combination)." Data description: date - date and time of measurements. year month - month of year coded 1-12 tod - time of day in half hour intervals coded 0-47 toy - time of year, as proportion. timeCount - cumulative time scaled to [0,1] load - grid load in megawatts. load48 - grid load 24 hours (48 half hours) previously. temp - average daily temperature (celsius) temp95 - exponentially smoothed lagged temperature dow - day of week