This repository contains a set of tools for performing end-to-end disaggregation of single-point energy signals at a low-resolution timescale of 15 - 30 minutes, including algorithms for supervised and unsupervised decomposition of signals into component appliance-level signals.
In particular, the repository contains:
- the
disaggregatormodule, a package of tools for an assortment of disaggregation tasks proto, a directory of ipython notebooks and scripts organized by topic.tests, a directory containing unit tests for thedisaggregatormodule.docs, a directory containing documentation including sphinx autogen directives and instructions for contributors.
Please view documentation at /docs/sphinx/html/index.html
Documentation can be found in the docs directory.
git clone https://github.com/dssg/wikienergy
In order to import the disaggregator package, at the wikienergy level type: python setup.py install
You may have to use the sudo command if a Permission Denied error occurs.
Tutorials can be found here
The proto directory contains a set of ipython notebooks with (less organized) analysis in greater depth.
To run the test suite, please refer to the developer guide.
If you are a developer working on this project via dssg, feel free to clone our website repo.
git clone https://github.com/dssg/smartenergyactions.org
We have our website running our disaggregation modules at smartenergyactions.org
Phil Ngo | Miguel Perez | Stephen Suffian | Sabina Tomkins | Matthew Gee | Varun Chandola
For a set of tools designed for evaluating and measuring the effectiveness of new algorithms, please instead use nilmtk (Non-Intrusive Load Monitoring Toolkit), which is designed with the needs of researchers in mind.