Is Coding needed for Data Science? - Role of Programming in Data Science

Data science is a combination of mathematics & statistics, programming and domain expertise. The rising penetration of high-speed internet has fueled the growth of people learning programming languages. 

So, in an era where coding is considered a life skill, one question arises: Will coding skills help in data science career? The short answer is 'Yes!', it'll help. The point to remember here is that it is only 40% of the task.
The rest is explaining the mathematical & statistical basis and findings to stakeholders and decision-makers. Domain expertise and mathematical & statistical understanding will help you along with coding skills.

Coding skills come in handy for data problems. A lot of basic things like data cleaning, data manipulation, loading libraries, etc. require programming knowledge. Good coding skills will help you circumvent the issues that initially come with data science problems. Many data scientists, regardless of their knowledge of the necessary steps required to solve the business problem, are not good at coding. So they face difficulty loading & manipulating the data, and getting the necessary libraries for implementing the desired solution.

Many data science problems can be solved with the help of libraries of programming languages like Python and R. If you are good at coding, you'll be able to troubleshoot the issues & problems very well, unlike those who are good at mathematics & statistics but don't have knowledge of loading packages, libraries, creating environments, etc. 

But when you present your solution to the end user, they generally ask very basic questions. These questions can be bucketed into two categories:

1. Domain-specific question

2. Mathematical & statistical assumptions of the solution

So, it is essential to understand the domain problem as well as the mathematics & statistics behind your solution. This puts you in a position to explain the solution to the end-user and stakeholders, it could be a data analyst or someone in a higher position in the company, like the person who is taking charge of sales or marketing. Understanding mathematics & statistics also helps in deciding the steps needed to solve a given problem. It also helps in determining which algorithm to prefer over another. 

Coding will help you to some extent. You still need a mathematics & statistics foundation and domain expertise. If you don't have one, you will have to gain that skill. 

If you ask me, I'll advise you to start with mathematical & statistical foundations. Domain expertise will come through experience. There is no substitute for experience in terms of domain expertise. But mathematics & statistics can be learned in a limited amount of time. It doesn't require 7-8 years of experience, unlike acquiring domain expertise. 9 months to 1.5 years are more than enough for you to master statistics & mathematics.

So, the conclusion of this post is that coding will help you a lot, but coding is not the entirety of data science.