Rule-Based Multidimensional Data Quality Assessment Using Contexts

It is an accepted fact that a value for a data quality metric can be acceptable or not, depending on the context in which data are produced and consumed. In particular, in a data warehouse (DW), the context for the value of a measure is given by the dimensions, and external data. In this paper we propose the use of logic rules to assess the quality of measures in a DW, accounting for the context in which these measures are considered. For this, we propose the use of three sets of rules: one, for representing the DW; a second one, for defining the particular context for the measures in the warehouse; and a third one for representing data quality metrics. This provides an uniform, elegant, and flexible framework for context-aware DW quality assessment. Our representation is implementation independent, and not only allows us to assess the quality of measures at the lowest granularity level in a data cube, but also the quality of aggregate and dimension data.