# timeshuffle

A lightweight privacy method based on time shuffling designed for sharing \
power profiles based data in open forums.

Conceptulised in the Energy and Power Group.
Energy and Power Group is part of the Department of Engineering Science,
University of Oxford.

A profile is split into periods with datapoints scrambled within each period.
The mean value of each period is maintained between the scrambled data and
the raw profile. For example, if the original
profile was a HH profile of 48 datapoints and periods was chosen as 8,
the profile would be shuffled within 8 periods, each containing 6
datapoints representing 3 hours. Each of those periods in the scrambled
profile will have the same mean value as the equivalent period in the
original profile. Therefore, if the scrambled profile is shared openly,
only 3 hourly data can be interpreted accurately.

Author: Scot Wheeler, [email protected]