As one of the fastest evolving areas of tech, data science has seen a rise up the corporate agenda as less and less leaders base business decisions on guess work. With added capabilities such as artificial intelligence (AI) and the edge complementing the work of data scientists, the field is becoming more accessible to employees, but this still requires training of data skills, on the most part. In this article, we explore some key emerging trends in data science, as believed by experts in the field.
Firstly, it’s believed that the involvement of AI and machine learning (ML) will increase further, and enable more industries to become truly data-centric.
“As businesses start to see the benefits of artificial intelligence and machine learning enabled platforms, they will invest in these technologies further,” said Douggie Melville-Clarke, head of data science at Duco.
“In fact, the Duco State of Reconciliation report – which surveyed 300 heads of global reconciliation utilities, including chief operating officers, heads of financial control and heads of finance transformation – found that 42% of those surveyed will investigate the use of more machine learning in 2021 for the purposes of intelligent data automation.”
Melville-Clarke went on to cite the insurance industry, often perceived as a sector that’s had difficulty innovating due to high levels of regulation, as an example for future success when it comes to data science.
He explained: “The insurance industry, for example, has already embraced automation for processes such as underwriting and quote generation. But the more valuable use of artificial intelligence and machine learning is to increase your service and market share through uses like constrained customisation.
“Personalisation is one of the key ways that banks and insurance companies can differentiate themselves, but without machine learning this can be a lengthy and expensive process.
“Machine learning can help these industries tailor their products to meet the individual consumers’ needs in a much more cost-effective way, bettering the customer experience and increasing customisation.”
Along with rising use of AI and ML models, organisations have been combining AI with robotic process automation (RPA), to reduce operational costs through automating decision making. This trend, known as hyperautomation, is predicted to help companies to continue innovating fast in a post-COVID environment in the next few years.
“In many ways, this isn’t a new concept — the key goal of enterprise investment in data science for the past decade has been to automate decision-making processes based on AI and ML,” explained Rich Pugh, co-founder and chief data scientist at Mango Solutions, an Ascent company.
“What is new here is that hyperautomation is underpinned by an ‘RPA-first’ approach that can turbocharge process automation and drive increased collaboration across analytic and IT functions.
“Business leaders need to focus on how to harness enterprise automation and continuous intelligence to elevate the customer experience. Whether that is embedding intelligent thinking into the processes that will drive more informed decision making, such as deploying automation around pricing decisions to deliver a more efficient and personalised service, or leveraging richer real-time customer insights in conjunction with automation to execute highly relevant offers and new services at speed.
“Embarking on the hyperautomation journey begins with achieving some realistic and measurable future outcomes. Specifically, this should include aiming for high-value processes, focusing on automation and change, and initiating a structure to gather the data that will enable future success.”
Dan Sommer, senior director at Qlik, identified software-as-a-service (SaaS) and a self-service approach among users, along with a shift in advanced analytics, as a notable emerging trend in data science.
“To those in the industry, it’s clear that SaaS will be everyone’s new best friend – with a greater migration of databases and applications from on premise to cloud environments,” said Sommer.
“Cloud computing has helped many businesses, organisations, and schools to keep the lights on in virtual environments – and we’re now going to see an enhanced focus on SaaS as hybrid operations look set to remain.
“In addition, we’ll see self-service evolving to self-sufficiency when it comes to effectively using data and analytics. Empowering users to access data, insights and business logic earlier and more intuitively will enable the move from visualisation self-service to data self-sufficiency in the near future.
“Finally, advanced analytics need to look different. In uncertain times, we can no longer count on backward-looking data to build a comprehensive model of the future. Instead, we need to give particular focus to, rather than exclude outliers – and this will define how we tackle threats going forward too.”
With employees gradually becoming more comfortable with using data science tools to make decisions, while aided by automation and machine intelligence, a concept that’s materialised as a hot topic for the next stage of development is the concept of ‘data fabric’.
Trevor Morgan, product manager at comforte AG, explained: “A data fabric is more of an architectural overlay on top of massive enterprise data ecosystems. The data fabric unifies disparate data sources and streams across many different topologies (both on-premise and in the cloud), and provides multiple ways of accessing and working with that data for organisational personnel, and with the larger fabric as a contextual backdrop.
“For large enterprises that are moving with hyper-agility while working with multiple or many Big Data environments, data fabric technology will provide the means to harness all this information and make it workable throughout the enterprise.”
Another important trend to consider regarding the future of data science is the new career paths and jobs that are set to emerge in the coming years.
“According to the World Economic Forum (WEF)’s Future of Job’s Report 2020, 94% of UK employers plan to hire new permanent staff with skills relevant to new technologies and expect existing employees to pick up new skills on the job,” said Anthony Tattersall, vice-president, enterprise, EMEA at Coursera.
“What’s more, WEF’s top emerging jobs in the UK — data scientists, AI and machine learning specialists, big data and Internet of Things — all call for skills of this nature.
“We therefore envision access to a variety of job-relevant credentials, including a path to entry-level digital jobs, will be key to reskilling at scale and accelerating economic recovery in the years ahead.”
In regards to new roles to emerge in data science, Adi Pendyala, senior director at Aspen Technology, predicts the emergence of the ‘Industrial Data Scientist’: “These scientists will be a new breed of tech-driven, data-empowered domain experts with access to more industrial data than ever before, as well as the accessible AI/ML and analytics tools needed to translate that information into actionable intelligence across the enterprise.
“Industrial data scientists will represent a new kind of crossroads between our traditional understanding of citizen data scientists and industrial domain experts: workers who possess the domain expertise of the latter but are increasingly shifting over to the data realm occupied by the former.”
Many organisations are being impacted by a shortage of data scientists in proportion to demand, but Julien Alteirac, regional vice-president, UK&I at Snowflake, believes that new tools, powered by ML, could help to mitigate this skills gap in the near future.
“When it comes to analysing data, most organisations employ an abundance of data analysts and a limited number of data scientists, due in large part to the limited supply and high costs associated with data scientists,” said Alteirac.
“Since analysts lack the data science expertise required to build ML models, data scientists have become a potential bottleneck for broadening the use of ML. However, new and improved ML tools which are more user-friendly are helping organisations realise the power of data science.
“Data analysts are empowered with access to powerful models without needing to manually build them. Specifically, automated machine learning (AutoML) and AI services via APIs are removing the need to manually prepare data and then build and train models. AutoML tools and AI services lower the barrier to entry for ML, so almost anyone will now be able to access and use data science without requiring an academic background.”