How Data Science Will Evolve in an Increasingly AI-Driven World
By Arijit Banerjee
Rapid advances in digital technologies, data, and analytics are transforming the role of data science in the future of work. While the popularity and excitement around data science is at an all-time high, how organisations value the data science discipline depends on whether they are digital-natives or non-natives. Most organisations have a dedicated data science team but giants like Netflix, Airbnb, Amazon, and Google, who claim to be completely data science driven companies employ data scientists in every business function. Companies who report phenomenal success with data often go a step further and employ another ace up their sleeves – Artificial Intelligence (AI). Combined with data science, AI presents an incredible opportunity to handle humongous amounts of data, derive meaningful insights, and enable intelligent decision making. But while data science as a career is the hottest one right now across the world, with demand far exceeding supply, this doesn’t take away from worries regarding its obsolescence in future. Nowadays, forward-looking youngsters are afraid to take the plunge into data science, thinking Machine Learning (ML) and AI will eat away their jobs in 20 years.
Fortunately, most of those concerns are pointless. What data scientists need to realize, instead, is that only routine data science functions such as data aggregation, collation, cleaning, and reporting will be increasingly automated and handled by robots in the future workplace. To remain relevant, data scientists will need to evolve in their roles by acquiring AI handling training to be able to tackle higher projects and drive value.
The new skills that data scientists need in an AI-driven world
Though programming, statistics and quantitative analysis still remain as must-have skills if one aspires to become a data scientist, knowledge of machine learning has become the new imperative on job. This is because AI has penetrated every ounce of data science and even data scientists who don’t implement ML models themselves, can benefit from learning the fundamentals of ML in order to create prototypes to test assumptions, select and create features, and identify areas of strength and opportunity in existing ML systems. Data science knowledge coupled with ML training is particularly sought after in fields such as statisticians, physicists, operations researchers, and others who have to develop top-down or bottoms-up models to address business problems.
The real work of a modern data scientist is to analyze and model the data
While knowledge of technical programming languages such as R and Python are undoubtedly important, a modern data scientist must acquire and continuously brush up his/her AI and machine learning skills because working with ML libraries and data visualization libraries is a key aspect of the job. The real work begins after data cleansing and involves a fair degree of data modeling, making basic (for entry level jobs) and advanced (for mid-senior positions) AI skills a critical imperative. The bottom line: traditional approaches towards data science no longer suffice in the modern world as they are both resource and time exhaustive, while ‘real-time’ is the new imperative for success. AI with its capability to fast track data interpretation and generate actionable insights, will offer data scientists a valuable method to move up the value chain and retain relevance in the future of work.