10 things every aspiring data scientist needs to know

The Harvard article “Data Scientist: The Sexiest Job of the 21st Century” first sparked my interest in the data science field. At the time, I had spent 3.5 years in management consulting and had built a great reputation building models and developing projections in MS Excel. After reading that article, I realized data science was a great intersection between my role at the time and programming (techy stuff) which I always had a desire to learn. This desire came from my strong interest in techy things while growing up. Thanks to my dad, I spent my formative years surrounded by all sorts of computers even when they weren’t so mainstream in the town I grew up.

Months after reading that article, I decided to join the newly formed data and analytics team in the company I work with. The decision seemed like a no brainer but it wasn’t the conventional thing to do because I was joining the team as a senior with no experience in the space. As a result of this, the initial months were a bit challenging but this spurred me to invest quality time into teaching myself data science.

I adopted the self-learning approach which started out as fun but later became time-consuming and challenging. I have put together 10 things I wish I was told when I started my data science journey a year ago.

I am sure someone out there definitely needs this!

1. Learn statistics and maths: If you ever failed statistics and maths, now is a good time to pick it up. You won’t get far as a data scientist without having a good understanding of statistical concepts, linear algebra and calculus. I know it gets complex at some point but for now, start with the basic stuff and keep building on that. Chances are that it’s not as tough as you think it is. And if you grew up in an African home you probably are conversant with the phrase below.

“The people who do it don’t have two heads”

Here are some courses I found helpful….

• Basic statistics by the University of Amsterdam
• Intro to descriptive statistics (Udacity)
• Intro to inferential statistics (Udacity)
• Khan Academy (for calculus and linear algebra)
• Blue1Brown (for calculus and linear algebra)

2. Learn to programme: It is easy to think programming is rocket science especially when you don’t have a tech background but it actually isn’t. There are several courses and articles that make it really easy to get started.Obviously, don’t expect to churn out super cool stuff from day one; it will never happen. But if you remain consistent, someday it will. R or Python? Well, I would leave that for you to decide.

I found data camp, code academy and solo learn for mobile extremely helpful. These would help you get started.

Other helpful resources:

• R in a nutshell by Joseph Adler
• R for Data Science by Garrett Grolemund and Hadley Wickham
• Python Data Science Handbook: Essential Tools for Working with Data by Jake VanderPlas
• Data Science from Scratch: First Principles with Python by Joel Grus

3. Study, study, practice, practice, and practice even more: You need to devote quality time to studying and practicing a whole lot. You can start with really simple stuff and build up on that. Just make sure you spend extensive time studying and practicing preferably daily. The field is really wide and technical so there are certain areas you need to read over and over again.Remember your study is never complete until you practice. See it as an iterative process. When you practice, more questions come up and then you are forced to study again. You would definitely learn a lot by reading and DOING. Always remember that data science is an applied field.

Helpful resources:

• An Intro to Statistical Learning by Trevor Hastie, Robert Tibshirani et al
• Analytics in a big data world by Bart Baesens
• Applied predictive analytics by Dean Abbott
• Applied predictive modeling by Max Kuhn

4. Stay hungry and curious: Curiosity doesn’t kill the data scientist; curiosity only kills the cat. So relax! As a data scientist, you need to stay hungry and curious to learn. In data science, there are so many concepts to learn and there are new ones flying around every day. You need to keep yourself abreast of changes and trends. Read books, articles, deconstruct codes just make sure your hunger never dies out. There may be times when you are reading one article and that same article provides about four links to various others. At that point don’t feel overwhelmed, just stay hungry. Remember, it comes with the job.

I have found towards data science, Analytics Vidhya and kdnuggets very helpful.

5. Put some structure to your learning: The concept of self-learning in data science always sounds really cool until you get into it and realize it is very time consuming and challenging. Data science is wide and it is easy to get lost when navigating through various areas which you need to learn. It is very important not to wander aimlessly around various topics in data science because it is harder to connect the dots that way. You need to structure your learning by taking courses and joining learning paths on various platforms.You don’t need to break the bank to do this. You can start with the free stuff first.

I found the following helpful

• Intro to data science (Udacity)
• Intro to Machine learning (Udacity)
• Kaggle learning paths
• Analytics Vidhya data science learning path

6. Join a community/meet up or get a mentor: The saying “together, everyone achieves much more” is definitely true in data science. In this journey, you need people. Never be a lone ranger. Chances are that you would burn out quickly and spend endless time getting to your destination. However with mentors and buddies, you could get really far. Don’t go weeks and months wasting time on something someone could have explained to you in less than a minute. Someone has probably done what you are trying to do, so don’t reinvent the wheel.

On the flipside, don’t be too quick to get help when you haven’t tried well enough. There are a lot of lessons you could learn from your own mistakes and research.

7. Get on Kaggle/competitions as soon as you can: Competitions tend to speed up the learning process. Like I always say, start with the simple stuff. The Titanic competition or the house price prediction competition on Kaggle are good starting points. Seeing yourself rise up that leaderboard is some good motivation. It gives you that feeling of progress; at least up until you hit a glass ceiling and find it hard to improve the accuracy of your model. Then the learning continues because you need to find out ways to improve that model. So get in competitions and share your codes for people to critique.

8. The best time is now: You are probably wondering if you lived all your life under a rock because you just found out about data science and you have read several stories about people who have been in this business for many years. Well, they say the best time to plant a tree was twenty years ago and the next best time is now. It’s never late. Quit whining about how you could have started data science a billion years ago. What’s most important is the fact that you started the journey and you are getting your hands dirty. You deserve a pat on the back for that.

9. Keep going, never give up: When the going gets tough, the tough get going. Don’t kid yourself, data science is no child’s play. It is going to get tough at some point. You need to stay focused and remain consistent. Celebrate those little milestones: learning your first algorithm, your first competition entry etc. Celebrate them.

10. Hang points 1–9 on a large frame on your wall and get your hands dirty. Always remember you are not alone so sit back and relax. It is definitely going to be a joyful ride.

Good luck :)

First published in towards data science publication on Medium.

Insightful? Leave a comment or thought...