Differentially private synthetic linked 2011 Census and mortality
Synthetic Census 2011 linked with death data, both created through statistical modelling to make it non-sensitive.
Synthetic data created by the Office for National Statistics (ONS) Data Science Campus. It is created by statistically modelling the original data (Census 2011 linked with mortality data) and then using those models to generate new data values that reproduce the original data statistical properties. Any biases already in the real data may propagate though to this synthetic version.
Created using a differentially private algorithm for categorical data synthesis. This algorithm seeks to preserve the most important statistical properties within the data while protecting the confidentiality of the data contributors, according to a mathematical definition of privacy. The privacy budget, epsilon equals 1.0.
This will be used in the Integrated Data Service (IDS) to facilitate analysis and innovation whilst maintaining the principle of data minimisation. It will also facilitate data access within the IDS, while preventing disclosure of confidential respondent information. This dataset is also known by Synthetic linked Census 2011 mortality. Data available only to provider approved projects.
All data hosted on the IDS are de-identified.