Energy-Forecasting

Data

The dataset contained energy production values for every hour of every day between 2020 and 2023 along with the source of production(solar or wind). The task was to train a model to be able to forecast production of energy

Features

1. Energy Source:
2. StartTime of Production: xx:xx:xx format
3. EndTime of Production: xx:xx:xx format
4. Date of Production
5. DayName of Production
6. DayNum(of Year) of Production
7. Month of Production

Process

After cleaning the data, pre-processing and engineering features, I trained a few basic models on the dataset to get a baseline for the RMSE. Then I trained an LSTM network on the data in tensorflow.
I've gone into depth of the decision making in the pre-processing and feature selection inside the src files