A dashboard for real-time tracking and 72-hour forecasting of US electricity demand using open-source tools.
USelectricity is an open-source project that provides real-time tracking and 72-hour forecasting of US electricity demand. It pulls data from the Energy Information Administration API, applies machine learning models for predictions, and displays results on an automated dashboard. The project solves the problem of monitoring and anticipating electricity consumption patterns using a fully automated, reproducible pipeline.
Data scientists and analysts interested in energy forecasting, time series modeling, and deploying data science projects with open-source tools. It's also valuable for developers learning about automation with GitHub Actions and Docker.
Developers choose USelectricity because it demonstrates a complete, production-ready data science workflow using free tools. Its unique selling point is the integration of data ingestion, forecasting, dashboarding, and automation into a single, transparent project.
Forecast the US demand for electricity
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Leverages GitHub Actions and Docker to refresh data hourly and regenerate forecasts every 72 hours, demonstrating a hands-off, production-ready setup from the README's automation component.
Built entirely with free tools like R, h2o, and jq, as emphasized in the philosophy, providing a cost-effective and transparent pipeline for data science deployment.
Pulls hourly electricity demand and generation data directly from the EIA API using R and jq, enabling current and historical insights for the lower 48 US states.
Serves as a comprehensive guide for learning data science workflows, covering everything from API ingestion to automated dashboard updates with flexdashboard.
Focuses solely on the lower 48 US states, making it unsuitable for global or regional energy monitoring without significant code modifications.
Relies on a Generalized Linear Model (GLM) from h2o, which may not capture complex time series patterns as effectively as more advanced algorithms like LSTMs or ensembles.
The README is marked as 'WIP', indicating that setup instructions and troubleshooting guides might be lacking or outdated, potentially hindering new users.
Heavily dependent on R and specific packages, which could pose integration challenges for teams accustomed to Python or other data science ecosystems.