So, this must be the 100th article that you are reading on how to become a data scientist. And, the fact that you are reading it, means you’ve still not figured out the right way to approach this field. So, what is so special about this article? Why should you read this one and not the others? Well, for starters if you are still reading this means you haven’t really found the right article to guide you.
The issue may be that you have learnt all the Machine Learning algorithms out there and still, no company is ready to give you a chance. Welcome to the club!
So, let’s take a step back and a moment to think about how does a recruiter or the person in power to give you this job thinks. First thing as a fresher in this field you should understand that knowing machine learning itself doesn’t keep your company afloat. The fact that you can create some kind of business value with the use of machine learning does. For example, as an end-user who has no idea about technology, I would not care whether you used a complex neural network or something as simple as KNN. I would just want the product to recommend correctly. Next thing, is to understand that if you come from a non-engineering background, something like B. Com or B.C.A etc. and you feel that this is a stumbling block and makes you a less attractive employee, you are right. Why would I as an employer want to pay some fresher 4-6 lakh from that kind of background? Why not just pay a person with a Btech degree, at least they would know how to code. So, how does someone cross these hurdles? Well, the answer is pretty simple: develop right skill sets and network with people.
CODE
Good for you if you know all the ML algorithms out there. But, if you don’t know how to code basic Fibonacci series or prime number finder, then what is the point of entering this field? You have to teach yourself the basic level of coding and learn how to optimize code or your software will crash miserably. Not even a fancy ML algorithm will be able to save it. So, teach yourself basic coding, understand when you should use if-else statement, for loops, or a while statement. Your fundamentals should be strong so that you don’t get uprooted. Learn how to do Web Scraping, as you will need it at some point.
MS – EXCEL
Something as simple as MS-Excel will help you become a data scientist? Yes, you read that right. In reality, you need MS-Excel often and more than you need those fancy algorithms. Your day-to-day job as a fresher will require you to be able to work your way through this simple tool. Nobody is going to directly let you work with fancy algorithms. So, learn some advance MS-Excel functions like V-Lookup, H-Lookup, Pivot Table, filtering and more. This will help you in the long run, too. And while you are at it learn how to automate MS-Excel via macros.
MATH
If you are starting as a Data Scientist, you should know basic math as all the algorithms are based on math. So, learn probability, statistics, linear algebra. You don’t have to go thoroughly in this area but you should know why some algorithms perform better than the others under certain conditions.
DATA WRANGLING
Understand that real-life data are going to be messy. You have to learn your way around these messy data-sets. Most of the times, they are going to be missing data or there might be some unexplained anomalies like outliers. You have to learn how to work with them. Learn how to format dates in a meaningful way, creating dummy variables or encoding them. Pre-processing your data in the right way is the key to some positive changes in your product.
SQL
Learn basic SQL to help yourself with the retrieval of data from servers or you will end up employing some expert for something so basic. As a data scientist, you need to know how to get that data and create your own tables for test cases. Learn basic SQL commands like SELECT, INSERT, UPDATE, CREATE, DROP, ALTER. You should be able to use basic functions like AVG, SUM etc. Do learn all the JOINs.
COMMUNICATION & VISUALIZATION
Learn how to make beautiful visualizations because, at the end of the day, you have to communicate your findings to a human. And, from what I know humans love stories including nerds like ourselves. So, familiarize yourself with data visualization tools like matplotlib, ggplot, plotly, or Tableau. And, take vanity in making some interactive visualizations for your team members and clients. Just learn one of these tools properly and you are good to go.
CERTIFICATES & PROJECTS
When you have learnt these skills, use them. Also, get some meaningful certificates to back your passion for Data Science along with meaningful projects. Create a project which just showcases you EDA (Exploratory Data Analysis) power, you don’t even have to use any fancy algorithm. Work with different kinds of datasets. Explore Time Series, Sentiment Analysis, and NLP- based projects . Participate in some Data Science competitions available online because it will help in your resume too.
Make sure you have the right kind of resume. Try to network and find yourself a mentor who is working in this field. Focus on developing these skills and keep learning constantly because this field is ever-changing and in the long run self-learning helps not the degree. Understand that companies want to hire someone who understands business and can build an end-to-end solution of any data science project at hand. Make yourself commercially viable, someone who when hired by a company will help increase revenue. They need you to be able to add value to their business and increase profits via machine learning. Starting small is the key. And, get into this field only if you are amazed by how data can provide you with answers to all the questions in this world. Don’t just do it for the money, because it will get boring and tiresome at the end as it is dynamic and you cannot be static.