This algorithm implements the geo-indistinguishability algorithm as outlined in [1]. Noise is generated on a 2D laplacian distribution based on the epsilon value set in the plugin settings and the protection distance set in the tool. The noise is generated and applied to each point individually within the dataset - i.e. it assumes that the location of each points is completely independent of the location of any other point within the dataset. As such spatial statistics will only be relatively accurate based on the privacy level set in the tool.
Any point layer may be used as input. Note that the calculation is done on a 2d plane in the projected units so care should be taken when looking at the projection of the incoming dataset.
An offset distance for the noise calculation. Approximately 59% of points will be within this distance of their origin.
If this parameter is checked then the offset distance will be limited to the distance of the 95% confidence interval. This does limit the protection guaranteed by the service but it guarantees there will be no large outliers in terms of point distance moved.
A dataset with all the original attributes of the original dataset with the points offset by a random amount. Note the offset of each point is independent so if the spatial autocorrelation of the data is important at the original resolution the utility may be lost using this tool.
The numerical value of the 95% confidence interval. As this is a number output this is not visible unless the tool is run from the command prompt or within a model.
One example of using this value would be to create buffers around the anonymized features to that you can supply an anonymized area hat should contain the input feature (for example, for heritage features).
[1] Andrés, M.E., Bordenabe, N.E., Chatzikokolakis, K., and Palamidessi, P. 2013. 'Geo-indistinguishability: Differential Privacy for Location-Based Systems', In the Proceedings of the 2013 ACM SIGSAC conference on Computer and Communications Security (CCS'13). New York, New York, USA: ACM Press, pp. 901-914. Online at http://arxiv.org/abs/1212.1984v3