Computer Laboratory

Gates Building Energy Use


Run Visualisation

Bug Reports

Units and Costs

Raw Data

Fridge Raw Data

Demo Code

To Do List

Previous Versions

Raw Data

The raw data used by the visualisation tool is directly available for download at

In addition, the raw data for fridge monitoring can be found at

Directory Structure

The data is organised in a structure which mimics the electricity distribution tree within the building. A given directory may have data files and/or subdirectories - for example:


is a directory that contains the readings for the lighting circuit on the first floor that comes off the CS1 riser.

File Names

Currently there are 2 types of data files, one with 2 minute time series (one per day) and one with hourly time series (one per month):

  • "S-m##-YYYY-MM-DD.json" - 2 minute time series.
  • "S-m##-YYYY-MM.json" - 1 hour time series.

For example, "S-m23-2011-05-01.json" would be the 2 minute series of meter S-m23 for the 1st of May 2001, and "S-m19-2011-07-22.json" would be the 1 hour series of meter S-m19 on the 22nd of July 2011.

Data Format

The data files are in JavaScript Object Notation (JSON). This is so that the files can be easily manipulated in JavaScript - indeed our visualisation tools are built on top of JavaScript packages that do most of the heavy lifting. You can find demo code to help you get started if you want to perform any analysis of the data here (Or you can always look at the source of the tool itself - after all you download it when you run it...)

Each file contains key-value pairs for identifying the data. The "data" value is time series of energy readings from the meter, presented as an array of [time stamp, power] pairs. The power is the average (in kilowatts) over the interval for the time stamp. The timestamps are milliseconds since the unix epoch (1 Jan 1970).

More keys will be added in the future. Current keys include:

  • "data" - 2D Array containing Integer timestamps and Floating Point power measurements in the form [[TIMESTAMP, POWER],[TIMESTAMP, POWER]].
  • "coverage" - String containing information of the coverage of the meter.
  • "label" - String identifying the meter number, and the month and year for the file in the form "## YYYY-MM".
  • "path" - String identifying the directory to which the data file belongs.
  • "description" - String containing information of what the electricity is used for, e.g. lighting, sockets, air conditioning..
  • "room" - String of the meter location.
  • "type" - String identifying the type of meter from which the data was obtained.
  • "ts" - String identifying the month and year for the file in the form "YYYY-MM".

We frequently regenerate the published data from the real raw data so new keys will be added to old time series, (and old ones may disappear - although at some point we will commit to backwards compatibility).