Correspondence analysis of massive data: key role of resolution scale of the analytics

Abstract

Resolution scales of analytics can be coupled in the following way. The principal or main analysis is carried out on a low resolution data encoding. This simply expresses
that data are aggregated. We consider where data aggregation is interpretationally of value; and where it is of computational benefit. Once the principal or main analysis is
carried out, using supplementary elements, we can locate or map rows or columns, i.e. individuals, or attributes or attribute modalities, in the semantic, factor space. From the
overall perspectives we proceed to address specific aspects of our data. While the above provides focus in analytical processing, what is so very important also is context. Here to
be described is how context gives rise to qualitative as well as quantitative effectiveness and impact assessment.