When Damian Bushong applied for his first data job, his resume was, in his own words, “frankly, anemic.”
He didn’t have any data-related work experience he could point to, but he had majored in computer science and had years of experience doing open-source development projects in his own time. So he included a link to his GitHub profile along with his resume, as a sort of data science portfolio.
His GitHub included three projects he had developed: one very small project, one mid-sized project and a framework. He specifically offered this variety of projects to demonstrate the versatility and breadth of his skills, as well as his attention to detail and consistency across projects.
It worked. Bushong got the job as a developer for UNCOMN, a business-to-business management and technology consulting services company. Today, he is an engineer on UNCOMN’s cloud team.
4 Tips to Make Your Data Science Portfolio Stand Out
- Include projects that embody your strengths and the variety of your skills. Think about projects you can discuss in depth.
- Highlight projects with real-world applications. Consider including Kaggle competitions or hypothetical projects if you don’t have much work experience to show.
- Communicate the problem, process and outcome of your projects. Include a blog or other writing samples that can give recruiters more insight into your strategy.
- Give a glimpse of who you are as a person. You could add an “about” page or include non-project-related blog posts that show your interests.
“Having those projects up and available on GitHub as part of a portfolio to demonstrate what I could do was instrumental to my hiring,” he said. “It was how I could show my skills, rather than tell the interviewer about them.”
A data science portfolio can be a great way to stand out to recruiters, particularly for people earlier in their careers. But if you choose to offer one, you want to make sure it works for you. An effective portfolio communicates both the thought process behind its projects and their creator — you — to catch a recruiter’s attention.
Here’s how to make sure yours does just that.
Focus on Projects With Real-World Applications
First off, be strategic about what projects to include. Showing off quality work means being selective. Less is usually more when it comes to choosing what to include, according to Gianna Driver, chief human resources officer at Exabeam, a cybersecurity management platform. Projects you select should be your best and they should be examples you are ready to talk about in depth with recruiters at an interview.
It is also important that the projects you showcase have a clear real-world application. For some, this might mean actual projects from past jobs (if you’re allowed to share those), but those aren’t the only options.
For those with little data work experience, Patrick Dever, founder of Coupon Ninja, an online coupon-finding platform, recommended candidates create and include targeted hypothetical projects. For example, if a job posting describes data visualization as a big part of the role, a hypothetical project for that posting should show off your data visualization skills. When doing so, you can target it to the sorts of business needs you know or expect the company has.
“The best data skills portfolios aren’t dependent on the type of platform, but rather the quality of application that a portfolio illustrates.”
“This type of personalization would require some research on the applicant’s part,” Dever said. “Still, it will be worth it, and the employer would appreciate that you paid attention to their needs and tried to make sure they know that you can handle those tasks.”
Including Kaggle competitions is another good strategy for those with little work experience and seasoned candidates alike. Bruce Martin, global talent acquisition director at Emburse, an expense management and application processing automation platform, said he likes to see candidates include their Kaggle competition standings and work in portfolios because the competitions ask participants to solve real-world challenges and are sponsored by corporations and research organizations.
When it comes to showcasing these examples, where you do it is less important. While Bushong had a great experience with using GitHub for his portfolio, many other sites serve a similar purpose, including Kaggle, IBM Data Science Community, Stack Overflow, Behance or even Reddit. Personal sites that you’ve created yourself can work too.
“The best data skills portfolios aren’t dependent on the type of platform, but rather the quality of application that a portfolio illustrates,” said Tina Hawk, senior vice president of HR at GoodHire, a digital employment background checks provider.
Write About Your Projects, Your Thought Process and Yourself
The best portfolios are those that give recruiters context. This can come in many forms: You can include write-ups with each project, link your blog to your portfolio or include a blog section if you are using a personally coded site as your portfolio.
“We love resumes and portfolios that communicate a problem the candidate was given, how they thought outside the box and about downstream business impact, and ultimately, what they created and the difference it made,” said Driver.
Writing about your projects also has another benefit, she said — it shows off your communication skills.
“Talent in a tech company must be able to work cross-functionally and communicate with non-technical talent,” Driver said. “If technical talent can effectively communicate their competency and the business impact of their expertise, we will most likely aggressively pursue a conversation with that candidate.”
Offering a data science portfolio is about showing off your skills, and by extension, you. Recruiters want to know about you. How do you think? What are your areas of interest? How do you solve problems?
Including an “about” page on your portfolio is a very direct way of communicating this. If you are including a blog in your portfolio, you might consider having non-project-oriented blog posts about your interests or hobbies to give potential employers a sense of who you are. Regardless of how you do it, the goal is to communicate to recruiters about your process and yourself.
Do You Always Need to Offer a Portfolio?
Short answer: no.
With few exceptions, like for more visual or design-based positions, recruiters looking to hire data science talent don’t expect portfolios included with resumes. They are nice things to have — icing on the cake, according to Driver.
But including a strategic data science portfolio can help you stand out from other applicants. This can be especially useful for those just getting into the field with little, if any, work experience to demonstrate their skills. It could also be helpful for more experienced candidates applying for a very desirable or otherwise highly competitive position.
Those who have more robust work experience they can detail on a resume, however, don’t necessarily need to include a data science portfolio, particularly in the current job market. And companies probably won’t ask for them or discount applicants who don’t include them, according to Dawn Mitchell, senior vice president of HR at Appian, an enterprise low-code workflow platform.
“To stay a top employer, you have to understand what candidates are looking for and you have to respond accordingly,” she said. “They’re not going to go through those obstacles that they might have gone through previously. The talent pool has just gotten tighter.”
Ian Jones, director of talent acquisition at Appian, called this an evolution of the tech jobs market. He also thinks the world of data and tech hiring may be moving past the point of needing — or even wanting — portfolios from experienced candidates.
Since many data and tech jobs involve proprietary code or other information that can’t be shared publicly, portfolios have to be filled with mostly side projects done outside of work. Asking for a portfolio can amount to asking for proof of working outside of work.
Jones reflected on his past experiences of hiring data and tech workers in the nineties and early 2000s when companies “almost required side projects.” But the focus has shifted recently, in his experience. Not only are talented data and tech workers — likely those with experience working at bootstrapping startups — no longer willing to work over 90-hour weeks only to go home and do more off-the-clock work-style projects, he said, but increasingly companies don’t want that either. Some are recognizing the danger of employee burnout and prioritizing the mental and emotional health of their employees as part of good business.
While the current job market might make portfolios unnecessary for some data workers, the field is evolving rapidly.
“New languages, frameworks, tools, methods, and more pop up on a regular basis,” said Bushong. “It’s difficult even for professionals in the field to evaluate skills with this constant churn of change.”
Data science portfolios are a way to communicate to recruiters what you are capable of in that, or any, context. If you choose to offer one, your portfolio is — or should be — all about making you stand out.