In an information technology landscape full of platforms, the ‘data science platform’ is now an official thing – but what is it, or what is one?
The curiously named John Snow Labs company is a specialist in Artificial Intelligence DataOps (that’s operations with a specific focus on data crunching) and its team thinks it can define what this term data science actually means.
The company says that professionals in healthcare, pharmaceutical, finance and other sectors (presumably legal is in there too) are using data science platforms when trying to extract factual information from long, free-text documents.
So can we pin down a more complete definition?
According to KD Nuggets, “A data science platform a cohesive software application that offers a mixture of building blocks used for creating many kinds of data science solutions, as well as drawing actionable insight for business processes, products and services. ”
According to Ali Naqvi, data lab product Manager at John Snow Labs, “A solid data science platform includes big data integration, data wrangling tools, data discovery and exploration, machine learning algorithms, experimentation tools, team collaboration tools and automated tools to deploy, test and monitor trained models in production. It will have a unified user interface, centralised administration and security controls — and an infrastructure that is highly robust and scalable to meet the demands of enterprise, business-critical needs.”
The firm emphasises that organisations should use data science platforms to create maturity and discipline around data science as an organisational capability, instead of only a technical skill held by a select few.
Markets and markets estimates that the data science platform market is estimated to grow from USD 19.58 billion in 2016 to USD 101.37 billion by 2021.
Magical analyst house Gartner now uses data science alongside machine learning to create one of its quadrants. As a piece of terminology in the ever-widening lexicon of IT, the data science platform has now come to be.