Opinion

The first concrete step that I took in my Data Science journey was registering to an online certificate. It was going smooth and easy. I was just following through and learning the basics. After some time, I started to think that it was easy to be a data scientist. I would just get a few of these certificates and start looking for jobs. I did collect those certificates in no time. Then guess what?
Welcome to the real world!
Certificates are good at teaching you the basics. They get you familiar with the field of data science. However, they are definitely not a game changer.
It is worth noting that I’m talking about the certificates received by completing MOOC courses. The ones that require you to pass a comprehensive exam like AWS certificates are on some other page.
Going back to our discussion, I have listed my 5 reasons why you should work on projects rather than collecting numerous certificates.
1. Certificates do not offer enough "active learning"

Certificates do not challenge the learners. If you pay attention to the lectures and read through the materials, you can easily pass the graded exams. They don’t push you hard enough to actually make you absorb the topics. If you don’t think hard enough and push yourself, the information will stay in the short-term memory and be forgotten.
As stated in this article by Nick Dam, "changes in neural connections, which are fundamental for learning to take place in the brain, do not seem to occur when learning experiences are not active. Many research studies suggest that active engagement is a prerequisite for changes in the brain. Not surprisingly, just listening to a presentation or lecture will not lead to learning."
Most of the time is spent on listening to lectures in MOOC courses. On the other hand, doing projects is a great example for active learning. They challenge you to actively engage in the entire process.
2. Not thinking outside the box

Certificates are designed to make you strictly follow a curriculum. It actually makes sense because you build up the knowledge from basic to advance. However, the map is drawn for you. You are not challenged to think outside the box. On the other side, there is no map or limit for the projects in the field of data science. You can define a problem, design a solution and implement it in any business domain. I think this is the most challenging part of being a data scientist. Data scientists use data to solve problems. In real world, neither problems nor data come to your plate.
It is, of course, important and needed to learn the tools and techniques that can be learned from certificates. However, if you cannot create value using these tools and techniques, what is the point of mastering the tools? To be able to create value, you need to think outside the box. Thinking about project ideas and actually implementing them build your path to create value with data science. The use cases discussed in the certificates are bounded by particular domains. However, the potential value of data science projects is much more. Just focusing on certificates prevents you from thinking outside the box.
3. Reality is full of surprises
Certificates only cover predictable scenarios. However, real life is full of surprises. It is not possible or feasible to cover every possible obstacle in a certificate. It also makes them too long to mention every possible solution or technique in detail. A too-long curriculum will bring your motivation down and thus completion ratios will be very low.
After I completed a few MOOC courses, I felt like a pandas master. I knew everything I needed to do data analysis with pandas. However, when I started my first project, I felt like I’m standing at the start point of a long road. Although it was a simple, introductory project, I spent so much time doing the following steps repetitively.
- I face a problem
- Start looking for a solution
- Learn lots of stuff while looking for a solution
- Find the solution
Projects do teach you a lot. More importantly, the joy you feel when the project is completed motivates you to start a new one in no time.
4. Data is not served to you on a silver platter
Most of the certificates have the data ready for you. The data is clean and nicely formatted. It sometimes contains some missing values which are structured in a way that can be handled easily. In some cases, you are asked to do basic preprocessing. Then, you are ready!
I don’t think that is what happens for most data scientists. You need to think all the way through the journey of data. Data scientist usually need to have answers for the following questions:
- What kind of data is needed?
- How can it be collected?
- How can it be cleaned and preprocessed?
When you are working on projects, a big part of the work is data preparation. You gain hands on experience of data collection and preparation which is a highly valuable skill for data scientists. You cannot improve that skill by collecting certificates.

5. Feasibility matters
Certificates walk you through the pipeline of a typical data science project from data collection to model evaluation. You can create a model for a sample use case. However, feasibility is not discussed deeply in those certificates. Feasibility is just as important as any other step of the project. If your solution is too complex to be implemented, then it is worthless. By doing projects, you also need to think about the feasibility. You not only try to build a model but also look for ways to deploy your model and create value out of it. You may have heard of the Netflix Prize competition on Kaggle with an award of $1 million. Netflix decided not to implement the winning model because it was not feasible. When you are doing independent projects, you don’t actually have to deploy your models or put them in production. However, you at least think through the feasibility.
Just like any learning process, we start doing simple projects and then level up by doing more and more. We eventually come to a point that our projects have the potential to be implemented in production. But, to be able to reach production level, we need to start doing projects. Collecting a bunch of certificates will not bring us to that point. I would recommend spending time on projects after learning the basics. The best practice is to get your hands dirty and do lots of projects.
Thank you for reading. Please let me know if you have any feedback.