Being a data scientist is the best. That’s not just a flippant perception, it’s literally the ‘best’ career according to a survey by job reviewing website Glassdoor.
The research took into account three key factors equally: earning potential, job satisfaction, and the number of job openings available. With a median salary of $110k a year, over 4,500 job vacancies in the US alone, and the freedom to wear Converse trainers to work – no questions asked – it’s fair to say that data scientists are loving life.
To be fair they deserve it. The decision to step into the world of analysing data in an era of information overload across every professional sector was a smart one. These scientists are able to extrude actionable insight from data at a time where information is power.
Data is growing at an exponential rate; in quantity it is set to double every two years, with a 50-fold growth predicted between 2010 and 2020.
In financial services alone the statistics are mind-blowing, and it is not unusual for a large investment bank to have between 20,000 to 30,000 servers these days. The issue is that once you have all this data you need to be able to use it to avoid it becoming an untapped resource.
So on the one hand the six-figure salary these data scientists are demanding is a reflection of their timely and unique skill set. On the other hand, it reflects the complete imbalance between supply and demand.
The need for these skillsets is skyrocketing across multiple industries far quicker than universities can produce oven-baked data science professionals. Research firm Gartner said in 2012 that there would be a shortage of 100,000 data scientists in the US by 2020. The European Commission then said in 2015 that up to 77% of data science jobs remained unfilled at the time and predicted a 160% increase in demand between by 2020.
According to IBM, the number of jobs for all US data professionals will increase by 364,000 openings to 2.7 million by 2020. Of those openings, 59% of all demand will be in finance and insurance, professional services, and IT.
A worrying shortage
After a very quick Google search it seemed that Columbia University had one of the top courses in data science around the world.
Rachel Fuld Cohen is a career development officer at the university’s data science institute and says that there has been a spike in applications to the masters programme.
“Our core courses, such as machine learning, algorithms, and exploratory data analysis and visualisation are receiving interest from students in programmes throughout the university,” she explains.
David Yao, a professor at Columbia’s department of industrial engineering and operations research, runs the centre for financial business analytics.
“The whole field of data science and data analytics is in huge demand,” he says. “That demand has changed our programme and our curriculum, and has shifted away from pricing.
“In New York City, suddenly all the universities have these programmes. Overall I would say data science is a relatively new field and we are only now seeing lots of these places setting up these programmes.”
The Harvard Business Review wrote in 2016 that the shortage of data scientists is becoming “a serious constraint” for certain sectors. Fortunately the number of students and courses are both expanding.
Part of the enticement might be that the IBM study found employers were willing to pay up to $8,736 above the median bachelor’s and graduate-level salaries for data scientists.
Can finance be fun?
So fingers crossed, the number of graduates is looking set to increase in the coming years, but for now the conversation remains that from the relatively small pool, do they want to come and work in custodian banks, or will the lure of companies like FAANG (Facebook, Amazon, Apple, Netflix and Google) be too tempting to ignore?
Michael Di Amore is a recent graduate with a masters degree in data science at Columbia University and says he actually finds finance more interesting due to these data challenge within finance.
“The idea that the financial time-series data has a low signal to noise ratio is widely known,” he explains. “This makes finance problems inherently difficult. I find problems dealing with alpha generation particularly appealing, as there are many competing forces in the space and everyone seems to have their own view and model of the world.
“To me, that is more interesting than common problems in the tech space such as Computer Vision or Recommender Systems.”
Global Custodian has written previously about the fight for talent just within the securities services industry alone, and the difficulties in attracting young talent to an industry very much stuck in its ways. The battle for data scientists will be no different.
Despite what Di Amore says about the work itself, the philosophy of a banking institution is entirely different from your typical Silicon Valley tech firm, despite a recent push to reform.
A change in culture – which most firms do seem to be currently working on – along with stimulating work and opportunities, seem to be at the core of graduates’ needs.
“Students are motivated by interesting problems and companies that have the proper infrastructure to support technical data science work,” explains Fuld Cohen. “If a financial institution invests in data science and makes data-related work and products core to their business, then students will want to work there.Mentorship and support at a first job is a huge draw as well.”
The securities services world is certainly creating more appealing roles for data scientists and other technology graduates. The banking world is looking to implement more sophisticated technologies into its operations in line with a shift towards focusing on digitisation and data analytics. Most custodian banks are now fully invested in innovation centres and artificial intelligence projects, enabling them to draw some of the aforementioned talent to their business.
Problems that need solving
The good news for students like Di Amore is that there is a huge opportunity to make a big difference to these major companies as they begin moving from proof of concept to real use cases.
“The most important requirement is that there are interesting problems to solve,” says Di Amore. “The second most important is that there’s actually data there, or some mechanism that can be developed to gather the data.”
Luckily for students such as Di Amore, there are plenty of problems in securities services industry which can be solved by data scientists.
BNP Paribas Securities Services has taken a two-fold approach in adding data scientists – recruiting straight from the best university for the subject in France, and training its staff internally. The latter is an approach that Cisco Systems is now famed for through its Data Science Certification programme, which is open to its employees of all backgrounds.
“We launched a training programme for those people with an engineering background to them up-skilled,” says Philippe Ruault, head of digital transformation at BNP Paribas.
“In my team I have two or three people attending courses on Saturdays and evenings at a polytechnique school.”
Columbia’s Yao still believes that financial services has pulling power when it comes to recruitment, but adds that the sector is still not one of the major users of data analytics.
“The financial services industry still attracts the best and brightest talents,” he says. “That kind of phenomenon is always there in financial services.
“The space is changing, it used to be long gruelling hours, but now places like Goldman Sachs are competing against the likes of Google and Facebook for similar talent.
“However, the biggest users are not in finance as of yet. Social science and medical science for example; their need for data science is still above that of the financial sector.”
Can’t we all just get along
As firms move from proof-of-concepts to real-life implementation with artificial intelligence, the input from the correct balance of resources is crucial for this development.
While data scientists are going to play a pivotal role, heads of business units have to be involved to contribute their expertise on the subject matter.
So the three key roles are the implementers, subject matter experts and data scientists. And all three of them need to be on the same page, which might be easier said than done.
“They need to work together so it needs to be organised, but it can be complicated to do,” explains Ruault.
“You have the people that know the businesses and know the legacy systems working with these engineers who can work on the use cases, and we need them to all be agile and able to work together. That includes the business owner needing to explain everything.”
There won’t be many heads of business units in the industry who are unwilling to embrace change on a public stage, with conferences and events full of veterans discussing the importance of blockchain, artificial intelligence and machine learning. However, when it comes to letting new blood into major decision-making roles, there will undoubtedly be tensions.
Keeping up morale
It’s not just banking institutions having to adapt to these changes but fund managers as well across the world, known even more for being stuck in their ways of doing things.
“You’ve got 40 to 50-year old portfolio managers with 20-year old data scientists,” says Larry Tabb, founder of research firm TABB Group. “Those older guys may not even let a junior analyst into the decision-making process and now you’ve got a data guy working on the strategy.”
Tabb believes a temporary separation from the business units might be beneficial while the data scientists become familiar with the procedures.
“The safest and easiest way to get them into becoming effective is to keep them in a separate area with a separate goal,” he says. “Don’t try to disperse them into day-to-day operations as much, get them working outside and building tools and services and then give support to all parts of the company.
“Keep them on their own floor, in their own area, in their own building. At least until they understand how it works, had successes and gained respect. They need to learn how it all works; and keep them independent. Without autonomy and the ability to seek success and solve problems and fairly quick successes, it can be very demoralising.”
Solving the mass shortage of data scientists in the open market is not an issue securities services firms can do much about, but becoming relevant and competitive enough to capture this talent is.
In this current climate, service providers are inundated with data that can be monetised if aggregated correctly. It will require custodians to incorporate AI or predictive analysis software into their systems to comb through the data and turn it into intelligible materials for clients to use. As such, custodians are likely to turn into providers of big data analytics, if of course, they can tap the most sought after people in the world.