In the last 10 years, the role of the data scientist has changed considerably. As a result, how a data scientist is hired has also changed.
Explaining how and who we hire in the analytics and data science space will answer one question that I often find raised by professionals, students and analytics aspirants while I am on a panel. The question: “What are the key skills that are expected of a data scientist/analyst?”
Rewind to 2005-2006. Analytics had just gone mainstream. Businesses started noticing the need for data analysts. At that time, an analyst was primarily viewed as someone who had a knowledge of statistics and was adept at number-crunching.
Enter the mathematicians, economists and statisticians. First, they were trained in the concepts and processes of business analytics, which meant imbibing relevant statistical concepts and then getting their hands on software — SAS was used predominantly used. Or, it was R/SPSS/Weka.
These newly-hired analysts would be trained in industry or sector-specific tasks. For example, building credit risk models for financial analysts and market mix models for marketing analysts. Anyone who did any form of analysis of numbers/ data would be called a data analyst or business analyst. The techniques they knew were specific to the domains they operated in.
Talent pool expanded
Fast forward to the turn of the decade, which saw a change in how analysts were perceived and what was expected of them. There seemed to be a mismatch between how analysts worked and how they needed to work.
The data presented by analysts was not clearly understood by business officers. There was a gap in communication between the analysts and the business professionals. Due to this, businesses began to appreciate the need for analysts who not only talked data, but were also able to bring some business context to that data.
With this, there was change in the hiring pattern.
The talent pool expanded to include MBA graduates, engineers, physicists and even psychologists.
Analytics was no longer the turf of only mathematicians and statisticians.
Anyone with domain knowledge in almost every sector could learn to apply techniques related to analytics and get insights from data for their respective business vertical.
There was a reversal of sorts in the hiring process — earlier analysts were hired and trained in business problems, and now business professionals were being trained in analytics.
This was the also the time that saw the rise of R as a strong and dependable alternative to SAS, and the implication of large volume of data, that is ‘Big Data’, was appreciated. Being an open source tool, R gained much popularity. Many small- and medium-sized companies were integrating analytics processes into their business strategy and the high price of SAS added to the increased adoption of R.
Due to the rise of Big Data, Hadoop was in demand, and so were related technologies used to store, manipulate, analyse, and transform the Big Data. I remember noticing how data scientists wanted to be differentiated from data analysts, and everybody coveted the title ‘Data Scientist’. But this boundary was still murky and there were no clear-cut definitions.
Multiple roles
Coming back to the present. We now see different job roles within analytics. There are different functions and job descriptions for business analysts, data scientists, big data analysts, machine learning specialists and data visualisers.
So, anyone setting her/his eyes on a career in analytics has a choice.
The time of an analytics generalist is over; now is the age of specialists. People with multiple skills are still preferred, but they should have a core analytics specialisation.
These classifications are still not mutually exclusive and are constantly evolving. We still do not have industry-wide acceptance of these individual job roles yet.
But what we do know is that eventually there will be. These newly emerging roles will undergo changes. Some of these might merge, and some might further split.
Machine learning and deep learning are already making waves, and this will enable analysis of a different kind of data.
Internet of Things will go further mainstream and so will the data generated by it. Data itself will guide in recruitment of data analysts and data scientists.
All of these advancements will require data scientists to constantly acquire newer skills, as there will never be an end to data; nor for the data scientist.
(Gaurav Vohra is CEO and co-founder of Jigsaw Academy)
COMMents
SHARE