Data, and how to work with it

Power is Work done divided by Time

Written by Phil Watt | 12-Mar-2024 00:58:43

Power is the rate, with respect to Time, at which Work is done. The legacy of my last name, Watt, has stayed with me over the decades. My ancestor, James Watt (probably, don’t check), was honoured in Western science by giving his last name to the unit of power. The name of my business is clumsily derived from this - 

Data Watt = Data Power = Data Joules per Second = DataJPS

Our new product, Relait, is inextricably linked to data power, and the ability to do useful work in a given time. We’re confident that, in data-intensive projects where you need to bring data together from different sources, Relait can save you around a third of your elapsed project time.

How? Data-intensive projects rely on assumptions about the data, and many of them do not hold true once the project starts. This is not about whether you have good or bad quality data. We believe that data quality is neither good nor bad. It just is. Knowing what it is, is vital to the success of your project. 

Knowing the quality of your data is not enough either - vital though it is. You must also know the relationships in your data. Not just inside one database or application, but within and across all your information domains in the enterprise. This is a hard problem to solve. 

AI predictions for this produce unreliable results. Traditional methods to discover relationships do not scale (joining candidate columns). We did the maths, and for a medium sized data warehouse, we reckon the compute time you need for 5,000 columns and 1TB data is about 15 million days - plus manual labour. This is why most businesses don’t bother.

This is where we come in. We have a unique, deterministic, scalable, and reliable way to uncover all of the relationships in your data. It won’t take 15 million days. It might take a week or so, to save you months of project work later on. Get in touch to find out more.