Strata

Managing interpretive data models

strata details.png

Two strata examples. The left one is one hypothetical of how we get to stratas of Government, Economy, Business ecology, Business type, Social norms/expectations, and specific business cultures. The right one is more abstract, showing one potential of how to develop strata from raw data.

Strata are everywhere, but we don’t often think of them. 

Each layer of strata is intended to communicate with adjacent strata. This is another place where two dimensions fail us: the strata aren’t actually stacked. The adjacencies aren’t in space, but in data. Strata use some of the same data to build a model, and that data can help people jump to another layer of strata. 

Strata are intended to help people find meaning in accumulated data.

As strata build on previous interpretations, data details often get lost in statistics. The 5% outliers still exist, but when population decisions are based on 95% adherence (which is huge!), 5% can look inconsequential. 

It’s rarely inconsequential to the outlier, and when the compiled datasets run to the millions, that’s tens of thousands of ‘outliers’.

Understanding a data set has strata can help us clue in that narrative truth is being promoted over perceptual truth or quality of truth. Modeling them can help us suss out where and how to navigate to an alternative view of the data.

Caveats: This is not robust and comprehensive – just an idea of what I’ve seen frequently. The taxonomy is a potential issue, and I think in this instance it's because the concept is still emerging. Some people instantly know what I’m talking about. Some people I expect to, don’t.

strata:
garbage-in, meaning