The role of data scientist started with the idea of the ‘unicorn’ who could do everything. It has evolved over the years to become a team sport rather than simply one person’s role.

The data scientist has been long prized with the financial world, seen as a boffin code-cracker equipped with the skills and know-how to unlock a wealth of insights trapped within vast data repositories.

Yet, almost as soon as their value was being properly realised, detractors were sounding the death knell for the role, pointing to a growing disconnect between what business heads expect from their data ‘nerve centres’ (a type of ‘magical’ and instantaneous value creator for the business) against the reality of a role that remains deeply rooted in the scientific method (one that often unearths more complex and deep-rooted organisational problems within the business).

Former Head of Data Strategy & Analytics at ME Bank and sought-after data consultant Dr Catherine Lopes, speaks with FST on why we must challenge our notions of what a data scientist can and should be to an FSI, as well as the evolution of the enterprise data strategy and avoiding the ethical minefields of customer data (mis)use.

FST Media: Relentless innovation is core to any digital-focused bank’s business agenda, with data largely underpinning this forward strategy.

As a data and analytics specialist, how do you feel a data and analytics roadmap can and should adapt to support wider business objectives?

Lopes: A digital-focused business ultimately has digital transformation at the centre of its focus. Digital transformation not only changes the way business value is delivered through digital channels to their customers, but it also transfers the data from ‘isolated’ to ‘integrated’ digitally. A well-interconnected data ecosystem forms the backbone on which all business strategies are executed.

A data and analytics roadmap is normally developed based on a combination of business objectives and technology constraints.

Data roadmaps on data governance, risk control, and cloud migration often require effort on a larger scale and over a longer period, and they are not going to get changed quickly.

 

In the meantime, the analytics roadmap – which includes automation, BI, advanced analytics, and AI – often evolves quickly over a short period due to the changes within the business agenda. Overall, a balance between steady data maturity development and a quick analytics insights delivery is the key to driving the organisation to innovate and transform.

FST Media: As organisations push to embrace newfangled digital innovations, particularly enabled through cloud, how can a data strategy best evolve in line with an organisation’s ongoing digital ambitions and progressive cloud adoption?

Lopes: Cloud adoption is always an important piece of the data strategy for any organisation that has ongoing digital ambitions. As mentioned previously, breaking siloed data and enabling a data ecosystem should be part of the overall target of any digital transformation.

Cloud transformation is the modern way of implementing data integration with the ease of having infrastructure management services in place. Many new features are only available in cloud-based data warehouses and platforms, which is another important reason for data teams to adopt cloud environments alongside digital. In the meantime, cloud adoption plays an important role in enabling organisations to take advantage of data science and AI development that can ensure the speed to delivery of business value.

FST Media: As the tide of data swells within organisations, questions remain as to how banks can and should utilise sensitive consumer data. While still advancing business objectives, how can banks ensure they utilise consumer data in an ethical and sound fashion?

Lopes: In my view, ethics should be the foundation, or indeed core values, of data and analytics. Data use should match the ethical guidelines within the organisation and across the industry or even broader.

Sensitive customer data needs to be treated with care, specifically with regards to the privacy, security, bias, and people impact.

To ensure well-rounded, ethical use of data, an organisation needs to manage it with a full lifecycle approach.

From the beginning of collecting data from customers, consent and transparency need to be implemented and provided. The organisation then needs to store and protect customers’ sensitive data with appropriate controls, encryption, transformation, and access. Crucially, business objectives need to be clearly articulated by the business division to end customers.

As part of the organisation’s DNA, an analytics solution development team needs to be trained with ethics practice guidelines for the data they use and the solutions they work on. This ensures ensure that, in advancing business objectives, they will not create ethical harm to customers based on the use or storage of their sensitive data.

FST Media: It’s no secret that sound data governance and literacy are key to building a competitive business. In your view, what does an effective data culture look like in the day-to-day functions of a business?

Lopes: In my view, a data culture needs to be embedded across a large segment of an organisation’s workforce; after all, if only a few people in an organisation are data literate, it’s hardly a culture at all.

A successful data culture is how effective people work, react, and think about data in their day-to-day job: a group of data champions who work with data and have the technical skills to perform data analytics tasks well; a group of business champions who recognise the value of using data to solve their problems and ask good questions of utilising their data properly; and a group of data leaders who provide guidance on how to use data and inspire more progress for the entire organisation.

FST Media: There appears a perplexing juxtaposition between those that advocate for data science as the ‘next big thing’ and those who are already predicting the role’s demise.

How is the role of data scientist evolving, what value does it hold for financial services, and why does go beyond simply ‘statistical analysis’?

Lopes: Indeed, I have observed the juxtaposition between those two views, which is not surprising given that we went or are still going through the hype of overselling data science on the market.

Data science is a stream of practice that includes data exploration, data preparation, model development, model evaluation, model deployment, business implementation and much more. The key to data science is that it uses more advanced and scientific methods, and indeed algorithms, to provide predictions and many other insights on solutions; it also overlaps with statistical analysis.

The role of data scientist started with the idea of the ‘unicorn’ who could do everything. It has evolved over the years to become a team sport rather than simply one person’s role.

 

This is one of the reasons for the ‘death’ path prediction, as these unicorn type data scientists are too hard to find or perhaps they don’t exist at all.

Despite the confusing juxtaposition, the use of data science has brought a lot of value to the financial industry, and many of their applications are disrupting the market already. Without data science methods, it will be extremely hard to implement personalisation or customised solutions on such a large scale with accuracy. Data science has also brought a lot of AI solutions home to the financial industry, such as document processing automation, voice recognition, and natural language processing that improve the operating efficiency and customer experience. Overall, I do see the financial services industry growing its maturity in data science.

FST Media: As an experienced corporate leader, what advice would you impart to aspiring data leaders in getting the most out of their teams and information assets?

Lopes: Data leaders need to be well equipped with a growth mindset, because it is a field moving at an incredible pace, matching the data volumes that have been accumulating exponentially in recent years. It is not simply a technology problem; it is also a strong teamwork and synchronisation problem to get data professionally managed. It requires strong business and operational acumen to bring business value out of data.

As with many other things, data or information assets cannot be built overnight. Instead, it takes many failures to evolve.

For example, we have seen many ”data products” that are derived from many repetitive ”data services”.

Again, I would like to address the importance of data ethics for data leaders. The impact of data ethics could be huge for the enterprise and data leaders need to bring people together with them on the journey. ◼

Dr Catherine Lopes was a featured speaker on FST’s Banking Digital Discussion – Unlocking the Business Value of Data & Tapping into the Power of Cloud and Data Visualisation panels last month.

She currently serves as a member on the AI, Data & Analytics & Intelligent Automation Networks board of advisors, as well as Director of Opsdo Analytics, a data and analytics consultant founded by Dr Lopes last year.