When we do data exploration, we have at our disposal a number of tools. These come with a variety of attributes and trade-offs, driven by the nature of the data, the complexity of the underlying phenomenon being analyzed, the sophistication of the user, and the effort the user is willing to expend.
At one end of the spectrum are highly automated tools with focus on certain representations of the data, such as Tableau, Spotfire, and QlikView. At the other are powerful statistical tools which leave much to the imagination, but demand more effort and attention to detail: R lattice/ggplot.googleviz, Python + Matplotlib, SAS (or WPS), and D3.js.
When we explore data, we do use the right tool for the right purpose. We keep all the options in our toolkit.
The end of data munging, data warehousing, systems for establishing data quality, and so-on is to aid decision-making. Many business decisions are and will probably always be made by humans. As such, at some point the data will touch the human mind.
We could break it out to multiple spectra, or provide more detail about what makes each special.