Name: The Reference Energy Disaggregation Data Set (REDD)

Data characteristics

Multivariate time series

Type of building(s)

Residential

Date

2011

Type of attributes

Real numbers

Number of buildings

6

Duration

weeks- to months

Level of details

Whole home, circuit level

Location of buildings

Massachusetts, US

Sample frequency

kHz/ a few Hz

Access to the dataset

The latest version of the data set is available at: http://redd.csail.mit.edu

Description of the data set

The REDD data set is presented in high frequency (kHz) and low frequency (Hz) groups. High frequency data include current and voltage measurements, while low frequency data cover Power measurement of individual circuits within the houses. The low frequency folder contains six folders, one for each house. Each sub-folder contains a number of channels for different circuits in the house. Labels of each channel which include appliance names are given in a label.dat file in each sub-folder.

Data set structure

The following file and directories are available from dataset

redd/

readme.txt

-- general information text

low freq/

-- ~1Hz power readings, whole home and circuits

high freq/

-- aligned and group current/voltage waveforms

high freq raw

-- raw current/voltage waveforms

Paper and citation:

J.Zico Kolter and Matthew J. Johnson. REDD: A public data set for energy disaggregation research. In proceedings of the SustKDD workshop on Data Mining Applications in Sustainability, 2011.