Should You Branch Into Data Science Now? My Unusual Answer.

, , No Comments
Nowadays, I see bankers, ICT engineers, project managers, sales managers, finance managers, HR managers, lawyers, accountants and many people already deep into a different career path all wanting to switch completely to data science.

I don't know where the pressure is coming from but some of them don't know that doing that switch can be retrogressive -- taking a lower grade/paying job and limiting your career growth. 

In the corporate world, a data scientist is HR managed as a data analyst using new set of tools. To a large extent, the same career paths await both. As a data scientist you will always work under a manager and it could the someone who is a project manager like you were before you made the switch. So that's the lower grade possibility I mentioned.

Also, your hiring is focused more on you doing a technical task more than on you doing a manager-level task, and it can be a very career limiting positioning. You are seen as data analyst with more powerful tools and new age skills. You will almost always have to answer more to other people and have (usually) no one to answer to you. You work back-end, the IT type of back office role.

If you do this (actual data science work) for too long (say 5 years), then there is a problem. Stagnancy. And that's because any significant career progression in a non-consulting company would mean you doing less of the actual data science work but managing other people to do them.

Let's call all what I have said - side 1 (of my opinion coin). I said it with the assumption that you do indeed learn at this your old age and limited free time to become a data scientist.

Side 2 is the expectation of how long short it would take them to be a proficient data analyst.

Taking a couple of courses will definitely help you stand out better than another candidate with same situation and ambition. But if I am looking to hire a data scientist, the skills I would be looking for is way beyond courses and training you have attended. I would be looking for real world project portfolios. I would be looking for your knowledge contribution in that space (deployed projects online, tutorials you post online, communities you are part of and active problem solving skills that demonstrate your mastery of the methods + tools). But then again, most companies in Nigeria are not really seeking Data Scientists. They are just looking for regular data analysts that can use more than Excel. Maybe Power BI, SQL and familiar with Python or R. And on the job you'll just end up using Excel and a BI tool. But as usual, they'll rather exaggerate the competence requirement. 

I see many people get that type of jobs and I see many of the bankers, ICT engineers, project managers, sales managers, finance managers, HR managers, lawyers and accountants actually think that is Data Science. 

Data Science. A Data Scientist is someone who would use more of python and R, build predictive models and deploy as products or integrated tools powering major processes within the company. Not someone doing daily and weekly reports with presentation and BI for management. 

In an insurance company, a data scientist would be building programmatic models that analyse risk and create on the fly customized insurance product for the insurance company's potential clients. So John can be on the insurance company's website and fill a well tuned form, and instead of being presented with static buckets of insurance covers he gets a cover fine-tuned for him with a premium cost that is unique to him (computed by a data powered dynamic algorithm). While the folks tasked with creating the day-to-day operations and finance report - regardless of the fancy tool they now use - are just data analysts.

In a lending or financial service company, a data scientist builds the model that predict default risk, do customer segmentation and do product offering simulation to determine perfect pricing on an ongoing basis.

In a telecoms company, a data scientist would model product offerings, do customer segmentation, do churn prediction model, build algorithms that power marketing activities and sentiment analysis.

The type of education and preparation to become a data scientist is nothing less than 9 months full-time (not spare-time) learning, online project contributions and real world projects building. And with a good (diploma level) grasp of statistics, algorithms, data structure and predictive modeling (machine learning). Anything less and you are not a data scientist.


Post a Comment

You can be sure of a response, a very relevant one too!

Click on Subscribe by Email just down below the comment box so you'll be notified of my response.