pillarxyz / wind-energy-generation-forecasting

Wind Energy Production Forecasting

In this project we use time series data about wind energy generation in germany to test and benchmark forecasting algorithms

Data

The data folder contains the two original datasets :

• time_series_60min_singleindex_filtered.csv (wind power generation)

• weather_data_filtered.csv (temperature data)

these datasets that were download from the OPSD Project are at an hourly rate from the start of 2017 till the end of 2019, we merged and resampled them into one dataset wind_temp_data.csv (daily)

Notebook

the notebook wind-energy-production-forecasting.ipynb contains all the analysis and feature engineering we've done on the data along with implementations of machine learning algorithms for forecasting like :

• ARIMAX
• Decision Trees
• Random Forests
• SVM
• ANN
• RNN
• LSTM

we also provided comparative metrics of all these algorithms at the end of the notebook