swissenergy-backend

🚀 live dashboard & detailed write-up 🚀

This repository contains the ML backend powering an energy consumption prediction dashboard.

Inspired by the SFOE's energy consumption dashboard, I figured it would be a great opportunity to talk about an end-to-end ML project, going over the challenges one encounters during

• Problem Understanding
• Data Ingestion
• Exploratory Data Analysis
• Machine Learning Modelling
• Industrialization
• Deployment

Important

I heavily encourage you to check out the 🚀 write-up 🚀 to make sense of this repo, as it goes through each stage methodically.

Note

The code for the frontend can be found here.

Repo structure

The repo is structured as follows

├── img/ 
├── model_server/ # ML backend
├── nb-dev/ # Notebooks created during the EDA/Modelling phase
├── tests/ # pytests
├── viz/ # Visualization built for the writeup
├── .gitignore 
├── .pre-commit-config.yaml 
├── Dockerfile
├── README.md
├── data_checks.ipynb # Used to manually check our data
├── compose.yml 
├── requirements.txt
└── sanity_checks.ipynb # Used to manually check our some inputs

Running the backend

The backend is meant to be run as a dockerized app, running off some machine. This project's write-up goes in depth about how to run the backend.