3.1 Themescope

We are assuming that the data you are working with comes from a context of great data diversity. For example, data from surveys or questionnaires, or from a database of clients, in which there is an abundance of different types of variables.

In these cases with a rich variety of variables, we can group those variables in themes. Each theme defining a point of view or multivariate reality. For instance, when we have a group of socio-economic variables, or when a set of variables have to do with preferences about a set of products. By refer to this approach as themescope, that is, a multidimensional description by themes.

The analysis strategy that we propose is to analyze individuals by themes. This involves selecting a particular theme in which the variables associated to it become the active variables. Having a group of active variables, we study the resemblances of the individuals according to this point of view. And then we add all the available information that has not been utilized, but that can shed some light in better understanding the relationship between individuals and variables, by using the projection of supplementary elements.

Projection of categorical supplementary variables

Figure 3.1: Projection of categorical supplementary variables

Various Perspectives, Diverse Analyses

When selecting a particular active theme, this does not stop us from selecting another theme that can then become active in itself. By changing active themes, we have a different perspective of the analyzed data, in analogous way to taking photos of the same subject from different angles.