Let's talk about portfolios
The word portfolio used to scare me. It sounded so professional and unattainable. I didn’t know what exactly was expected of me. I thought I would have to make my projects look very sleek and perfect. I thought it would have to include projects with some sort of breakthrough or a novel idea in data science. I thought that I would first have to be working professionally as a data scientist to have a portfolio. I thought maybe if I didn’t do projects for a client or in a professional setting, they wouldn’t matter. And it didn’t make it better that most advice on the internet about finding a job mentioned having a portfolio.
I’m pretty sure, as an aspiring data scientist, you hear about the importance of having a portfolio every other day. But there is no reason to be anxious about it. Let’s talk about portfolios.
A portfolio is nothing more than a collection of things you’ve worked on
Don’t be overwhelmed by the word itself. People like using fancy words and portfolio is a fancy word for saying a collection of your works. It is a collection of things you worked on, you put effort and thought into. Anything that illustrates what you learned so far and clever solutions you thought of.
Your work doesn’t have to be professional. There doesn’t have to be stakeholders. Think about it from an employer’s point of view. Before they hire someone, they want to be able to see with some evidence, what the applicant knows. Python might be one of the skills you mention on your CV, but can you show a piece of your code in Python? Your cover letter might say that you finished five online courses but can you show a project where you analysed the data and trained a machine learning algorithm yourself?
Employers mostly just want to see some evidence. Your work doesn’t have to be cutting-edge or perfect. It just needs to be somewhere where they can see it and look through it.
How can you make a portfolio?
All you have to do is to choose a platform where you can effectively present your work. A couple of popular platforms are: a personal website/blog, a GitHub account, a Medium blog.
No matter the format or the platform you choose to present your work, you need to cover a couple of points. Talk about
- The problem you decided to solve with your project,
- The data that you used,
- The problems with the data and how you solved them,
- The results you got and what they mean for your problem.
You can include charts and graphics but it really doesn’t have to be cosmetically perfect for your portfolio to matter. The content is what matters.
At the end of the day, you want to show that you are aware of main data science concepts, you know how to deal with problems, you know how to avoid potential pitfalls and you can think of creative solutions for potential problems.
Don't get stressed about amazing looking portfolios people put out there. Comparing yourself to people who already have a couple of years of professional experience is not useful. It’s much easier to collect projects once you’re working. You’ll get there. The first thing is to actually get a job and to get a job, you need to focus on learning and showing off what you learned.
Just focus on getting things out there
And you know what the best portfolio is? The one that exists. The important thing is to get your hands dirty. This is not new advice from me. I really think that the earlier you start doing projects the better. Because that’s how you build upon the theoretical knowledge you gained from online courses and books.
Trial and error takes time though. It might be somewhat inefficient to start a project and struggle your way through it by yourself. It helps to have someone around you who can help you make good choices on decision points, point out your mistakes and hint at what you should learn next.
I didn’t have anyone like that when I first ventured into data science. So it took me quite a bit of time to figure out how to make a proper data science project, what to look out for, how to decide on algorithms and many other things… Luckily, I had an internship that gave me enough space and time to learn things by trial and error. Not everyone might have an abundance of time to start their own project though.
I'm working on something that's going to help you with this exact problem. With it, you’ll be able to see how a data science project is done from start to finish and be ready to start your own projects that will form your portfolio. I will announce it in more detail soon. Stay tuned for updates!
In the meantime, think of things that you enjoy doing, watching or reading about. Can you come up with a project about any of them? If you can, write them down and when you’re ready, you can make a project about one of your interests to show off your skills to your next employer.