3 career paths in Data Science and why I chose the one less traveled?

When you talk about data science what comes to your mind?

Deep Learning, Machine Learning, and complex algorithms governing the tech giants. But, the reality is very few people will actually work with ML algorithms & even fewer will actually enjoy the process.

In fact, for most Data Analyses you will be fetching data from the right sources & simply summarizing through Power BI/ Tableau/Excel/PowerPoint. And that's a huge help for the majority of business decisions.

Not that you can't be an ML engineer or a data scientist - the question is which one adds value to businesses and goes well with your personality as well?

Before, even jumping into the colossal ocean of data analysis, you need at least some direction to start. You can always change the path if needed, but if you don't choose a direction, you will end up with a bunch of online courses meaning nothing.

Unfortunately, that's most people end up today.

Here's a way you can look at things

3 Approaches to learning data analysis:-

1)???Reporting

• ETL
• Pulling Data & Data Visualization
• Machine Learning Code (just the code not the math)

2)???Math

• Math behind Machine Learning/Deep Learning (stats & math)

3)???Domain & Communication

• Understanding the industry & business fundamental
• Story Telling – explaining your analysis to the stakeholders in business language

No matter where you start from,?you will have to go through the first stage. You will?have to understand the basics of Excel, SQL, Python, etc. to pull the right data.

But, there will be a stage where you will acquire working knowledge of this part.

The question is where do you stop? Do you want to master coding or take up something entirely different?

A.???You see, everyone wants to do the first part i.e. coding - try to look for Python courses, and you will find numerous courses and articles on the internet. Talk to any data analyst about his skills, and he'll talk about the technologies he knows.

B.????A few analysts want to understand the math behind the algorithms. E.g.: You might be able to apply K-Means clustering in python, to create customer segments. But do you know how that algorithm worked in the back end to create those segments?

C.????A tiny number of people want to acquire domain knowledge, improve communication & presentation skills, and cogently communicate the analysis to business folks.

So, try to connect the dots - look at kids, the younger they are the better they are at tech, because of the information advantage Gen Z has over Gen Y. Consequently, the coding part of data science will be filled with candidates good at programming. And, as the time passes their quantity will increase and quality will improve.

The Math guys will usually be in short supply since it’s a brain-draining exercise to create models, and understand the math behind them. Run a model a zillion times, until you achieve accuracy.

Finally, people who understand business and can communicate the findings to the end-users will be in extreme dearth. Not because, it is extremely difficult, but because no one wants to do it.

Because, most people want to talk about tech, a few about math, and even fewer about domain & communication.

Why don't people work on domain knowledge & communication skills?

Reason 1 - Who enters data science & why?

Data Science is a field of nerds - people willing to bury their faces in computers join it.

Now, if you ask them to maintain relations with stakeholders, understand requirements, present reports, and take feedback – they almost abhor it. A lot of Computer Science guys don’t like client-facing roles (my personal experience, might be wrong).

Even, if you are a bit extroverted - the data science work pattern forces you to bang your head on the computer for hours together with extreme focus. So, introversion is a by-product of this profession.

In fact, a big reason why some IT folks don't progress is their resistance to leadership roles. They are happy writing code for the back-end but don't want to lead people or manage clients.

Reason 2 - Coding is cocaine

Coding is an addictive activity in itself.

When you bang your head on a particular problem for nights together, you are finally able to build something, you get a sense of achievement. It's like solving a SUDOKU puzzle - the intellectual stimulation is addictive.

You don't want the headache of continuous calls with stakeholders or the stupidities of freshers reporting to you.

You love your terminal and pages-long code. So, are not willing to take up managerial roles.

Reason 3 - Not dying without it

Understanding business is not directly related to your work as much as coding is. For a lot of entry-level data analyst roles, if you understand SQL, a visualization tool, and basic ML, you will be able to execute stakeholders' orders, even without understanding business in depth.

You don't need in-depth business knowledge to create a SQL query & plot data in a Tableau Report, as per someone's instructions. And your senior/manager can always explain this report to the stakeholders.

To sum up:- People don't want to acquire domain knowledge & communication skills because they hate people facing roles, they love coding and they don't urgently need it.

Demand-Supply Gap

Because of the above-mentioned factors, there will always be a demand-supply gap

A natural outcome of this concept is, that as you move downwards in the list, the supply of people reduces.

Of course, to be a great Data Analyst you need to be good at the fundamentals of coding, math, and business. So, you will be the jack of all, and a master of?

You can’t master everything, right?

For instance, I am good at SQL, Power Bi, Excel, Business Understanding & Presentation Skills. But, for excellence, I have chosen a particular field of Data Science i.e. business understanding and communication. This will be my specialization.

Does it mean I don’t understand other aspects? Nope, I am madly in love with machine learning, and so I am learning various algorithms, the math behind them, and the data preparation part. I am also continuously upgrading SQL & Power Bi knowledge as well.

Because, if communication is the tongue of a data scientist, machine learning is his head, and SQL & Excel are his hands. Despite knowing everything about machine learning and math, he can’t do anything without a good understanding of machine learning & data pulling.

Having said that, I want to be known as someone who is brilliant at understanding business problems, brainstorming a solution, and communicating the results to business users.

Summary

Data Analysis is an enormous ocean with multiple disciplines. If you randomly start doing online courses, you will end up doing nothing. The best way is to pick up a direction, be it reporting, machine learning, coding, or any specific direction. Look at your nature, look at the market's future demand-supply scenario and keep on changing your direction until you find the best fit.

In general, the supply of techie nerds will be higher than demand, math folks will be in short supply & people with a good understanding of data + domain + communication will always be in dire need in the industry.

Have you chosen what area you want to excel in? feel free to mention it in the comments section.