I work on projects with clients, and help them understand the data they have, and what kind of business value it has. Most of today's companies already have systems that collect huge amounts of data about their operations, their users and their performance but do not know how to use such data to make business decisions and policies.

I first understand their business goals and translate their data into information, from which new insights can be gathered. On a typical day, I spend half the time analysing the data, and the remaining time communicating the insights I can find from them to the business executives.

The three main pillars of Data Science are

1. Statistics, including probability concepts, linear algebra and calculus. If you need a background on linear algebra and calculus, I recommend Khan Academy's Linear Algebra, Probability & Statistics, Multivariable Calculus and Optimization courses.

2. Computer science, meaning programming. It is very rare to see a data scientist who does not code. Today's languages of data science such as Python and R try to make the work of a data scientist easier. For getting started with Python, you may want to try this website called www.datacamp.com. It offers an interactive browser where you can learn Python for data science, R for data science, etc. For those interested in learning machine learning, Introduction to Statistical Learning is a very good starting point. For more advanced courses, Andrew Ng's Machine Learning and Deep Learning specialisations are very good and are highly recommended. But, to take this, you need to have a good hold on Python and calculus.

3. Business acumen - While business acumen can be hard to gather in the days you are starting, it is important to be always kept in mind. Without an understanding of what the business wants, the other skill sets are not valuable. For those with little or no experience in the data science field, a good method is to practice telling a data story. Imagine you find some interesting pattern in the data, as you learn. Ask yourself, “Am I able to communicate what this data is saying to a business user, and is it aligned to what this company values and cares most about?”

I like working with numbers, and the story they tell us. Data has become the asset of many companies today, and one of the best ways to unearth such assets is through Data Science. Like any profession, it is not always easy, but one that is highly satisfying if you have the right skillsets. I got inspired by the quote “In God we trust; all others must bring data” by Edward Demming, a renowned American statistician. I wanted to prove that I can unearth some insights from the data which is being supplied to me, and work towards unravelling the mysteries it presents. Also, the term 'Data Scientist' is quite newly coined, but the work data scientists did, has always existed. It used to be previously called 'statistician', 'quantitative analyst', etc. but it lacked the coolness factor it has today! :)

As I mentioned, data is the asset of all organisations today. So, mobility is not an issue, as the skills are transferrable between companies. But the catch is that, while statistics and programming skills are easier to transfer, what works in one business domain might not always work in the other. So, it is very important to have that business acumen when making a pivot in one's career. What works in the E-commerce industry might not work well in the F&B industry. Hence, I would recommend first understanding the business you will be involved with, and then focus on transferring the other skillsets. Because, even if you are technically very good with mathematics and are a good programmer, in the new industry, your findings can be deviating from what the domain needs. So, if you have a clear focus industry that you like to work on, try to get an internship in such companies. To do that, express some interest through some pet projects, showing that you have the skillset to perform.

Teamwork and communication would be the top two skills. These skills are very important as a data scientist typically will be working in a team of business analysts (one who understands the business use case, and plans the processes) and data engineers (one who will get the data from different sources). In most cases, data scientists need to explain what their algorithm does, or how they have used the data to gather insights in a manner that is not mathematical or code driven. This is a very important skill set, as not all business users understand all the algorithms. Communicating and defending what your work provides is the topmost skillset one needs to have as a Data Scientist. Coding skills in Python and R, which are today's languages of data science are a minimum. To get started, you can use Python more, which is fast growing into the industry standard for doing data science. Statistics and quantitative skills, the knack of handling numbers, are going to be necessary. Also, when communicating it is essential to make some simple graphs, and stories from your data, hence data visualisation skills will be relevant as well.