What we in data can learn from natural science
Soeren Kier Christensen, CEO.
Over the past three hundred years, the philosophy of Natural Science (namely, the idea of analysing the behaviour of natural processes and then constructing and using mathematical models to describe the results and make predictions about the outcome of additional observations) has been very successful.
Over the last ten decades, this idea has been extended into social sciences and economics but, unfortunately, here it has, on many occasions, been somewhat turned on its head. Mathematical models have been proposed as suitable for explaining social or economic behaviour, without the crucial activity of first verifying whether the assumptions implied in the models were likely to be valid foundations for the underlying human behavioural phenomena.
Time and again this has led to much effort being spent on trying to find elusive circumstances in the real world under which the model’s assumptions might be at least remotely or temporarily connected to reality, rather than ensuring from the outset that the scope of applicability of the models was well understood.
This deviation from the original idea of constructing models that fit all currently available data at the point of inception, has proven significantly less fruitful than its cousin in the natural sciences. Over the last two decades in particular, efforts have reverted towards the original inspiration with the advent of behavioural economics and behavioural psychology and the associated attempts at modelling group behaviour on the basis of observable traits of their individuals.
In many aspects of life, whether business, public administration, medicine, sports or elsewhere, the past two decades have brought a wealth of digital data which can similarly be used to build and test predictive data models of the behaviour or susceptibilities of customers, consumers, patients and other segments of the population.
However, this data often resides in multiple distinct and disjointed systems, where it is represented in data sources with a structure and nomenclature that makes it difficult to create a coherent and transparent unification of the data.
Finworks Data provides an environment in which business analysts and business experts can easily and readily create a laboratory where they can discover, analyse, test and automate the production of the data nuggets and the models that transform their data into actionable and intelligent insight.
This allows our customers to replicate the same methodology that has been so startlingly successful in natural sciences to deliver rapid and substantial facts-based insights into their own business.
Our unique data analysis and data modelling products and services have helped generate these insights in many different industries and activities.