A Challenge to Spatial Data Analysts:
Can You Analyze Morphed and Ribboned Data?
Abstract
We present some spatial data transformations that may be applied to sensitive
data to protect the data by disguising the true locations of points.
Transformed spatial data cannot be analyzed in the same way as un-transformed
data. In transforming spatial data, we must lose some metric properties;
and we
may choose to sacrifice some contiguity properties as well. Our goal is to
balance what we give up to protect confidentiality with what we keep to be
able
to do analysis. To illustrate spatial transformations that alter locations,
but preserve some ability to do spatial analysis, we cut a map with
scissors to
produce a single continuous closed strip of quadrilaterals (cuts remove some
contiguity relations). We then show how to transform our cyclic sequence of
quadrilaterals into a constant-width ribbon by an area-preserving
transformation. Finally, we show that some spatial analysis of relative point
densities may still be done on the transformed space. We then challenge the
audience to extend the paradigm!