The evolving role of tech in data science
New technology will have a huge influence on the evolving role of the data scientist, as more organisations invest heavily in data-driven products and services.
This is according to David Logan, modelling solution scientist at Absa, speaking to ITWeb on the side-lines of ITWeb Business Intelligence and Analytics Summit 2019, in Johannesburg this week.
Discussing the evolution of data science andits dependence on technology, Logan explained that data scientists contribute a wide range of skills to an organisation, with the main three being business operational skills, technical skills and statistical skills, with innovative use of technology playing a key role in all three.
"Data science in its current iteration is a natural progression of what people have been doing for ages. In the past, data science was known as decision science, and then it became information management. A few years ago, it was known as business intelligence. As the role of a data scientist evolves, the technology they use has almost exploded in a way which has meant some hard decisions have to be made.
"These decisions relate to deciding which technology to choose and use to perform certain tasks, while complying with regulatory requirements and sticking to the available budget. Decisions have consequences, which may impact on how you perform your work, and weighing the pros and cons of using certain technologies is important over and above reaching your goals."
The evolving role of a data scientist requires the use of analytics tools, technologies and languages to help data professionals extract insights and value from data, he continued.
"These technologies include big data platforms, data virtualisation and predictive tools, such as SQL Server Data Mining, SPSS Statistics Software, Python and an array of other tools."
When big data began entering the business world, most organisations were ill-equipped to deal with the vast amounts of data they were sitting on; hence growing expertise within the data science field has been a natural progression, particularly in the banking sector, he added.
"The regulatory framework has changed extensively over the years, such as the extent to which companies have to protect customer data. The majority of data that banks incorporate today come from two main areas; one is the internal system and the other from external sources such as credit scores, customer profiles. So it's important that banks consume this data in a more responsible manner."
Impact of automation
As technology advances in the data science field, many organisations across the globe are experimenting with artificial intelligence (AI) tools, which potentially work faster than data professionals, at a more affordable price.
A McKinsey report estimates that about 64-69% of the total time spent on data collection and processing can be automated.
The report notes AI can be applied to certain tasks typically performed by data engineers, such as preparing data, cleansing, checking for correctness and identifying outliers.
As certain roles became replaced by tech, Logan pointed out that new roles will be created.
"Perhaps some of the data engineering jobs may become absolute as they are easily automatable. But having that grounding there enables them to move up the value chain. Most of the data science jobs that have been replaced till now are the repetitive and mechanical jobs.
"The digital era means we will have to actively round up the experience and identify talent in all three areas by taking a data engineer and actively educating them with skills of a business analyst or a statistician or an econometrician.
"In this case, you are cross-skilling them with skills of the other two legs, to get them closer to the business, while retaining their original skills, instead of letting them continue down the same track and eventually phasing out their role."