Dimensional ecology in information architecture
So far I’ve aligned the basics of information structure to the marks (nodes and connections), form (juxtaposition and placement), and shading (information structures) that define an object in two-dimensional art, building towards a three-dimensional interpretation. Color and detail (encapsulated rich data) make it recognizable, and context (levels) provides narrative. A representational piece of art also includes perspective. Perspective sets a contextualized object in the world, providing not only more context but a deeper narrative. Perspective can make a flower monstrous, or disquieting, or lovely.
I consider perspective at least as important in information architecture. It helps us understand where and why information isn’t included in a particular model — and that it still exists and is relevant somewhere, somehow.
When we start talking about truly huge data stores, information structures, software and product ecologies, the only way to truly make sense of them is to acknowledge strata.
One of the big concepts in memory is chunking. The theory is that we’re recoding smaller units of information into larger, familiar units to help us fit what we know into finite working memory. We’re not losing the details, we’re just clumping them together for a while and putting them in storage. Basically, we’re creating categorical structures on the fly to keep our cognitive load at a reasonable level.
Strata are a type of chunking. They are bigger than levels, and can contain levels. They are as complex as any one dimensional model can be, and can contain innumerable useful single-dimension models to help understand what’s going on inside it.
Strata help people find meaning in accumulated data, specifically according to their needs. Strata helps us make broader decisions, more fluidly. It doesn’t make sense to have strata until there is complexity and depth to information.
I like to use ‘strata’ as the taxonomy for this concept because it implies layers and the potential for separation even while the edges might be fuzzy. A while ago I was hunting for where I came up with the word, pinging contacts who worked with big data, data visualizations, in information; all to try to figure out where I got it. Some people instantly knew what I was talking about. Some people I expected to recognize the concept, didn’t. Some people recognized the word in context. No one had a different word for me.
Then I looked up into a plains state sky when a big thunderstorm was rolling out. oh, yeah. Big clouds butting up against each other. A deluge of rain still beating against the windshield. The bright, actinic sun breaking through and limning the thunderclouds in clear definition of the stray white clouds following behind, and feathery clouds far, far above — layered in the sky. The sun also highlighted how the rain was falling here and there, but not over there. I started thinking about the abstracts of the sheer size and form of space, transition, and unseen interiors with the detail of millions of raindrops formed from even smaller pieces, to fall from the sky due to their accumulated, gravity-bound weight.
Here, also, is the biggest trick of information. It’s not one-dimensional. It’s not even two-dimensional, easily sketched on a page and encompassing everything. Try to go back and mentally visualize all the dimensions of π mentioned in encapsulated rich data into a single object, and you’ll start understanding data isn’t even three dimensional.
Data is multidimensional, bigger than my mind can visualize even with rolling thunderstorms in a huge prairie sky as a model; even with decades of artistic visualization, including three dimensional carved object making; even with the active imagination built on more decades of reading and seeing everything I read like a movie in my mind. Strata helps me chunk information into more manageable complexities and start teasing apart the necessary details. Then I can use critical thinking, more conversations, questions, and the willingness to be wrong to make sure it’s all hanging together, balanced and whole…at least enough to do what needs doing next, or soon.
Each layer of strata is intended to communicate with adjacent strata. A big sky never really ends, but simply rolls beyond our sight; the same happens with data and how it shares into various information structures. It doesn’t stop existing, it doesn’t cut off; it just rolls out of sight.
This concept applies to big, unweildy, complex, and interdependent information. Think ecologies of software, the more connected the more relevant (Atlassian, Google, Adobe, Microsoft). Think deep research repositories (medicine, chemistry). Think research repositories that require inference (astrophysics, philosophy). Think all the languages we use, how some can share alphabets, taxonomy, structure, and/or cadence, but can also be wholly separate.
This concept applies to how different business units use the same data in variant forms and computations to understand business health in the macro and micro forms. Think big picture (KPI’s, profit/loss, EBITDA) and team metrics (deadlines, units produced, hours invested).
This concept applies to navigation. A data analyst needs to navigate to the finest granularity, usually along multiple dimensions. A team might use part of that data to help make decisions, but they need a different set of navigation points for it to be useful and intuitive. A manager who needs to share information with other managers to start to make broader decisions needs a different set data yet, but maybe with access to granular data to build novel insights and suss out hiccups.
This is a place where two dimensions fail us: the strata aren’t actually stacked. The adjacencies aren’t in defined space, but nebulous data (unmarked in the below model) with uncountable potential connections in a vast and mostly undocumented connectome. Strata use some of the same data to build a model, and that data can help people jump to another layer of strata.
Consider, also, that this representation is built from a business hierarchy. This is just for a hierarchical structure, the easiest one to understand the data movements. In other words, as strata start being useful, perspective matters.
Understanding that information has strata can help us clue in to more complex realities and truths. Modeling them can help us suss out where and how to navigate decisions that better balances the whole.
Truly, the strata I nod to above is not even the full complexity. We can keep chunking them over and over again, even if we stick to the hierarchical structures embedded in so much of our information transference mechanisms, methodically ignoring the fuzzy edges and rolling-sky attributes of the underlying data stores.
This is one pieces of Why learn the fundamentals of information architecture structures?
