Why is scale an effective descriptor for data quality? The physical and ontological rationale for imprecision and level of detail

Abstract

Observations and processing of data create data and their quality. Quantitative descriptors of data quality must be justified by the properties of the observation process. In this contribution two unavoidable sources of imperfections in the observation of physical properties are identified and their influences on data collections analyzed. These are, firstly, the random noise disturbing precise measurements; secondly, finiteness of observations—only a finite number of observations is possible and each of it averages properties over an extended area. These two unavoidable imperfections of the data collection process determine data quality. Rational data quality measures must be derived from them: Precision is the effect of noise in the measurement. The finiteness of observations leads to a novel formalized and quantifiable approach to level of detail. The customary description of a geographic data set by ‘scale’ seems to relate these two sources of imperfection in a single characteristic; the theory described here justifies this approach for static representation of geographic space and shows how to extend it for spatio-temporal data.