Assitan Koné
Oct 16

How can you break into data science?

Table of contents


Assitan Koné
Founder @Codistwa
Empty space, drag to resize



Maybe you’ve been trying to get into data science but you’ve been feeling overwhelmed and even worse, you think that data science is not for you because you don’t have the level, and/or because you are bad at math.

If you want to succeed, getting started with data science should be a no-brainer. Let me guide you with these simple tips.

Set up your ultimate goal

You will need to create multiple goals, but what can really help you is to create your ultimate goal. yes, it’s a carrot, but this is what will help you stay motivated when things get hard. You will be outside your comfort zone multiple times (congratulations on that, I’m proud of you!), so the game must be worth the candle.

Maybe you have your “why”: a better job, high paying, stimulating intellectually, with the possibility to work remotely, whenever is important for you. But the “what” will be a very specific title. You can be tempted to learn everything that it’s possible in Data Science. But it’s not realistic at all. Even, in your lifetime.

This field is huge. Think about it. If you want to learn a musical instrument, you choose what it seems the most comfortable to you. For example piano, guitar, or violin. Then you learn the fundamentals and you practice this over and over this particular instrument, you don’t try to master every musical instrument. It’s the same thing for Data Science. You choose a path. And this is actually your success path, which can help you reach your ultimate goal.
So which path should you choose? The easiest way is to pick up a field or industry. So healthcare, finance, etc., or you can decide to do NLP, Computer Vision, or Time series. Now you may wonder how you can choose one. When you learn the fundamentals, of course, you can try do to different types of projects to find your preference. And when you find it, stick to it. Maybe you’ll see quickly that you love working on NLP projects, that you are interested in doing chatbots.

Organize your data science study

Now that you have your ultimate goal. It’s important to break down into smaller goals. Let's say you think that your ultimate goal will be achieved in one year, so 12 months. Now you can set your 90-day, monthly, and weekly goals.

You can decide to create a personal project in 90 days for example, or to get job interviews. In one month you can study an algorithm or technique. It’s better to set weekly goals because you don’t study too much during the week, so not daily, so you can stay motivated in the long run. For instance, you can decide to study 2 hours a week at the beginning.

You can break down your sessions using the Pomodoro technique, and for further, use tools like Toggl to track your time and Forrest to keep your distractions out.

And finally, you can use Notion to plan your goals.

Focus on the data science fundamentals

Your goal is not to do a PhD, as it’s better to work in the industry, because the pay is better and you’ll get the possibility to see your code in production. Trust me, I had considered doing a Ph.D. because I was so passionate about algorithms, but I realized that I wanted to help aspiring data scientists create data science projects for my business, and make an impact in the world. Maybe that’s your case too.

That being said, yes, it’s important to know the fundamentals but stick to what is related to machine learning. For example, statistics is a big field. Your goal is not to be a statistician but a data scientist. Meaning, that you will use some statistics but not all. That’s good news, right? Save time and be very specific and intentional when it comes to your learning.

If you feel the need to learn linear algebra, statistics, probabilities, and calculus, study them to develop a math intuition and to understand the next steps, which are machine learning algorithms. But don’t study these topics to apply to a PhD, in other words, don’t push too much. Just go on a free platform like Khan Academy and it will be okay.

Also, you don’t need to learn everything at the beginning of your journey. That is normal to do a back-and-forth between what you know and your gaps. Usually, we need to see something several times to really grasp the information. Now, if you don’t know how to code, I really encourage you to start now, because it takes time to get programming reflexes.

Now, how can you apply that to reach your data science goal? You can use the Pareto Principle. So, instead of focusing on 100% of the official data science roadmap, make sure that you work very hard on 20% which is your success path as I said before. It will be so much easier, and more agreeable.

Create your data science personal project

You will see many data science projects that I call “classic data science projects” like the Titanic, house price, etc. These projects are important to grasp the fundamental techniques. So you can definitively do these projects to improve your data science skills.

Then, if you really want to stand out, it’s important to create your own dataset, to explore it very well, so you can explain it to anyone, and adapt your speech to your audience. You can use jargon when talking to other data scientists, a hiring manager, or if you write a technical article, but make sure to have this habit of using simple words when explaining complex concepts, like if you were talking to a 10-year-old.

Write your awesome label here.

Design Your Custom Machine Learning Chatbot

Write your awesome label here.
This quiz aims to help you create a machine learning-powered chatbot tailored to your interests, passions, culture, values, and expertise area. Answer the following questions honestly to uncover your ideal chatbot concept.
Sign up. Be inspired. Code.

Get a FREE Machine Learning Roadmap!

Subscribe to our newsletter to get your gift.

Get tips to teach yourself data science without being overwelmed in your email box. Get secrets to think and act like a Data Scientist on a daily basis. 
Write your awesome label here.
Created with