The take-away

• Data scientists are working on new DNA sequencing technologies. They could create patient profiles that identify mutations and come up with diagnoses and treatments.
• The job has three parts: data engineering, decision-making, and entrepreneurship.

Soon there will not be enough superlatives to describe the profession of data scientist. Described by LinkedIn France as one of the “most sought after” jobs, it is also the “sexiest job of the 21^st century”, according to the Harvard Business Review, and quite simply the “best job” in America for the website Glassdoor. The profession’s appeal reflects its high salaries and plentiful job opportunities. But money and security alone do not make it a true vocation.

Data science is hardly new. In the 19^th century, scientists analysed marine logbooks and maternity-ward registers to solve problems and build knowledge. With the explosion of free information, the field has taken off. Yanlei Diao, professor at École Polytechnique in Paris, defines her job as “extracting intelligence from a vast amount of data, with the least possible human involvement”

“Making the world a better place”

Diao doesn’t know if her job is the sexiest, but it has meaning. “It’s about taking a powerful, effective approach,” she says. “It helps us understand our world, overcome limitations and make it a better place.”

A Berkeley PhD, Diao describes one of her projects involving genomic data analysis. “We want to revolutionise healthcare by making genomic analysis an integral part of medical care,” she explains. New DNA sequencing technology can be used to produce billions of fragments, but computers are needed to process all that data and reconstruct an individual’s genome. Data science is applied to create patient profiles that identify mutations and potentially come up with diagnoses and treatments. “It used to be that only rich people had access to this slow, costly technology,” she says. “Our goal is to make it available to everyone, with highly reliable results.”

An “exciting” job

Mastering the job takes a lot of personal dedication. “You have to be passionate and stay informed of new developments”, says Peter Wittwer, responsible for infrastructure at the Digital Communications department of the German multinational Siemens. “The fast pace makes it an exciting job.”

Taming big data requires a variety of skills, says Arjan van den Born, academic director of the Jheronimus Academy of Data Science in the Netherlands, co-founded by Eindhoven University of Technology and Tilburg University. These aptitudes depend on the position that the data scientist holds in the value chain. He defines three levels: data engineering, decision-making based on big data and data entrepreneurship. On the first level, the expert builds, tests and maintains databases and data-processing systems. The data engineer then transforms data into useful information through different techniques: machine learning, analytics, statistical methods or image recognition software. The second level requires data visualisation and communication skills to submit results that help people make better decisions.

Creating value

 Some data experts focus on what van den Born calls data entrepreneurship. “This third level involves implementing business models that use data or algorithms to increase revenue,” he explains. “An example would be adjusting ticket fares in line with demand.” These data scientists work at consulting firms and large companies, or may be self-employed.

Security and ethics

Because the profession gets to the very core of organisations, security and ethics are crucial. “I’m responsible for data hosting, data security and the resources provided to do that”, explains Wittwer. “I have to find a balance between data security and accessibility. If there are leaks, damage to the company’s image or financial position can be serious.”

Van den Born emphasises the importance of ethics in today’s world: “This is the time when we need to ask questions like, ‘Should algorithms be transparent?’ or ‘What are we legally allowed to do with data?’”

Ideal qualifications

In high demand on the job market, the position of data scientist is at the crossroads of several traditional academic disciplines.

• An undergraduate degree in applied mathematics, statistics or computer science from a technical university is a good start.
• A master’s degree in data science or big-data analytics is useful for positioning on the job market, or further training. Anyone geared towards entrepreneurship or who wants a good understanding of economics can also opt for a dual curriculum between an engineering and a business school.
• A PhD in data science opens doors in the industry. “Many PhDs are taken on at large companies such as Google and Facebook”, says Yanlei Diao of École Polytechnique. “They’re hired for their technical expertise but also because they’ve learnt to solve a problem that hasn’t been solved yet. Recruiters believe they’ll be able to solve lots of other problems during their careers.”