isaacflower/resilient_project_ts_forecasting

Notebooks for a workshop for the RESILIENT Project that looks at modelling mini-grids in Kenya # Resilient Project: Timeseries Forecasting Workshop Notebooks for a workshop for the RESILIENT Project that looks at modelling mini-grids in Kenya. These notebooks closely follow the following TensorFlow tutorials: - Neural Networks - Bayesian Structural Time Series I have simply applied some of the tutorial material to the Kenyan mini-grid data. ## Package Versions OS: Windows 10 Enterprise python==3.10.8 matplotlib==3.6.2 numpy==1.24.1 pandas==1.5.2 scikit-learn==1.2.1 seaborn==0.12.2 tensorflow==2.11.0 tensorflow-probability==0.19.0 ## Notebooks This repository contains 4 notebooks that work through the creation of `neural network` and `structural time series` forecasting models for `electrical load` and `PV generation` data. ## Accompanying Material I have uploaded the slides that were presented in the workshop (`slides.pdf`) and the accompanying workshop manual (`manual.pdf`). ## Exercises 1. Install Python through Anaconda: anaconda.com 2. Install the above packages (install tensorflow with pip) 3. Download the data and notebooks and try to run through the code. 4. Train your own neural network models to forecast the "panama_data" Rough template solutions are included in the `exercise_solutions.ipynb` notebook