Customer Logins

Obtain the data you need to make the most informed decisions by accessing our extensive portfolio of information, analytics, and expertise. Sign in to the product or service center of your choice.

Customer Logins

Data science: at the heart of business transformation

28 June 2018

Data science is not new. But the availability of new tools and technologies is changing the role of the data scientist, enabling them to produce the insights needed to deliver commercial value and effect business change.

Data scientists have been described as the rock stars of the financial world - highly sought-after and earning big bucks. Data is an essential component of any business, but with so much data available these days, only skilled data scientists can make any real sense of it and put it to commercial use.

Traditional statisticians and data analysts are increasingly regarded as too limited in their abilities and remit. They are being replaced by data scientists who combine statistical know-how with computer coding skills to create deeper insights on business data.

Kate Land, Chief Data Scientist and co-founder of Havelock London, a new investment management firm, and who has a PhD in cosmology from Oxford University, says data science is a more advanced discipline than statistics or data analysis and more suited to tackling business problems.

"Data scientists work out how to use data effectively in an organization," she says. "They understand technology and coding, maths and statistics, and also the business domain they are in, so they know the challenges facing the business and how value can be added.

"The components are not new. What is new is that data sets have got larger. Computers have become a bigger part of the story. Also, what's new is the term. I didn't identify as a data scientist a few years ago. The term has come out of a recognition of the importance of the activity." The role of the data scientist involves a lot more data processing to generate meaningful insight than the roles that preceded it.

Yaacov Mutnikas, Chief Data Scientist at IHS Markit, lists some of the methods he and other data scientists use: "Data mining, descriptive statistics, statistical regression, Monte Carlo techniques, stochastic processes and transfer learning. It's a lot."

Yet the science is still developing. "In time, new techniques will evolve," he says. "Data will get bigger, algorithms will be smarter, and the questions they have to answer will be more sophisticated."

Stephen Roberts, RAEng/Man Professor of Machine Learning at Oxford University, argues controversially that data science is not as scientific as it should be. "Data sets have become so large that we have come to rely on algorithms to sift through beyond-human-scale data sets, which means we no longer have a scientific audit trail of reasoning," he says. "We have incredibly fragile, irreproducible, non-verifiable algorithms that work, but which tomorrow could break catastrophically."

He says what is largely practiced is not true data science but "human- supervised algorithm development". For him, there must be an audit trail of reasoning, "which is the logic and principles of mathematics", and that is largely missing from what people call data science.

There are others who challenge the concept of data science. The statistician and writer Nate Silver in a speech to the American Statistical Association a few years ago said the job title "data scientist is a sexed-up term for a statistician" and that "people shouldn't berate the term statistician".

Eddie Bell, Head of Machine Learning at Ravelin, the fraud detection company, understands that point of view, but argues the two jobs are different because of advances in computing. A data scientist has to handle much more data than a traditional statistician ever did, and has the computing power and machine learning to deal with it.

"A data scientist is a statistician with a Mac," says Mr Bell. "The rise of data science came with the rise of machine learning, and although machine learning uses statistical techniques, it is distinct from statistics.

"The main difference between statistical analysis-minded practitioners and data scientists is about what they produce. An analyst looks at data and produces some objective insight. A data scientist is an engineer who produces not just insight, but also changes the business."

Mr Mutnikas agrees, but if data scientists are to be instrumental in transforming the organization, "the firm needs to have a strategic long-term view of how the data science fits into the fabric of its culture, and where it makes decisions based on data". He adds: "Senior management must make a commitment to acquire, curate and look at the data in a responsible way; only then can data scientists deliver productive insights that drive the business."


Michael Imeson is Senior Content Editor, Financial Times Live; and Contributing Editor, The Banker. This article is based on a panel session at "The Data and Disruptive Technology Forum: Rethinking the Financial Services Ecosystem" organised by the Financial Times and IHS Markit in London in June 2018.

IHS Markit provides industry-leading data, software and technology platforms and managed services to tackle some of the most difficult challenges in financial markets. We help our customers better understand complicated markets, reduce risk, operate more efficiently and comply with financial regulation.

Follow Financial Services



Follow Us

Filter Sort