Why We Explore

I am interacting with data, and yet I am not searching for anything in particular. I might not care should I not find anything of interest.

And yet I have a desire to ascribe a purpose to what I am doing, to give it a name is to give it value, it proves that my time can be accounted for.

If I call what I am doing “exploring”, what is my goal state? Do I want an overview of the data, a map of its high-level structure? What if the data is devoid of structure? Do I still need a map?

Perhaps I am not exploring, but merely preparing the data for analysis that I foresee occurring in the future. I am coordinating, sorting, aggregating, summarizing, and generating descriptive statistics that describe my data. I am simplifying and cleaning the data, I am plotting subsets of the data, I am arranging and annotating these plots.

If I don’t have a goal in mind, these activities don’t appear to be very pragmatic. I could do these activities because I assure myself that any emergent analytical question will be facilitated from these manipulations of the data. I do these activities because I want to refine my own conception of the data, my mental model. Interacting with the data externally facilitates the formation of this model in a way that would be difficult to accomplish internally in my mind. I can’t aggregate, sort, and coordinate all at once (I, like most of us, have a hard time keeping track of 7±2 bits of data in my head). External interactions with data makes the mental model tangible, less ephemeral and error-prone. I can then take my mental model and share it with you, whereupon it becomes your mental model too. A model of the data that we share captures our individual cognition, distributed across space and time.

This is an optimistic way of explaining how I interact with data: I appear organized, prepared, and ready to collaborate. But could there exist a simpler explanation, one that is based in self-interest, habit, or even boredom? Interacting with data can be stimulating, even in the absence of an analytical goal or a knowledge gap needing to be filled. What if I just want to be entertained? What if organizing, aggregating, and sorting data are merely diversions that make me feel powerful, giving me a sense of accomplishment that I have imposed a structure on something previously less structured. Or perhaps I’m just genuinely curious. I could also be seeking inspiration, or inspiration might be a byproduct of this diversion; I might identify a knowledge gap that might bring about an analytical goal, an actual need to understand the data by means of directed, pragmatic inquiry.

I wonder, will anyone care about this alternative explanation? Outside of casual settings, is there value in casting the act of interacting with data as a diversion, as entertainment? Data analysts in professional and academic settings are paid to be pragmatic, to analyze data with tangible goals, to use expensive data analysis tools intended for this purpose. This is likely to be an overgeneralization; I expect that many analysts are hired on the basis of their ability to solve problems in creative ways. So where does creativity come from? A murky question, no doubt, but nevertheless semantically linked to inspiration, diversion, and entertainment.

Further reading:

Kirsh, D. (2006). Distributed cognition: A methodological note. Pragmatics & Cognition, 249-262.

Liu, Z., Nersessian, N. and Stasko, J. T. (2008). Distributed cognition as a theoretical framework for information visualization. IEEE Trans. Visualization and Computer Graphics (TVCG) 14, 1173-80.

Toms, E. G. (2000). Understanding and facilitating the browsing of electronic text. International Journal of Human-Computer Studies 52, 423-452.

Leave a Reply

Your email address will not be published. Required fields are marked *