The truth is infinitely complex and a model is merely an
approximation to the truth. If the approximation is poor or misleading, then the
model is useless. ― T. Tarpey†
Every data model tells a story about the world. Behind every regression, every neural
network, every forecast is a set of assumptions – choices we make about what matters, what
can be ignored, and how the underlying reality behaves. These assumptions form the invisible
architecture of the insights we derive.
Data centricity is about getting that architecture right. Our focus goes beyond
collecting more data or cleaning it better. We ask: are the assumptions
that connect data to insights sound? When assumptions hold, models reveal useful insights.
When they don't, even the most complex models are useless.
At Data Centricity Lab, we study this gap – the space between raw data and reliable
insights for decision making. Our work spans the assumptions embedded in statistical methods
(parametric, nonparametric, semi-parametric), the assumptions required by different modeling
objectives (causal inference, prediction), and the assumptions that carry forward into the
decisions these models inform.
