How To Build A Data Science Portfolio Without An Analytics Job

Data practitioners from QuantumBlack share their tips on how to build a compelling data science portfolio before you even submit your first job application.

Many people seeking their first job often find themselves in an age-old and challenging loop: you can’t gain employment without experience and you can’t gain experience without employment. A similar refrain is common in the advanced analytics industry, where demonstrable analytics skills are valued even for entry-level roles. But how can you build a compelling data science portfolio before you’ve even taken part in your first professional project?

At QuantumBlack, we’re surrounded by practitioners who entered the data industry through a variety of traditional and non-traditional routes. We gathered insights from three colleagues across data science and data engineering roles who all acquired initial analytics skills ahead of entering the professional industry. They offer their advice below on how aspiring data engineers and data scientists can make a start on a portfolio before they’ve even submitted their first analytics job application.

Understand The Road You Want To Follow

The analytics industry may appear to have a high bar of entry and it’s common to meet colleagues with extensive years of study behind them. Despite this, it can be a fairly accessible industry. Practical knowledge is ultimately what candidates are assessed on, rather than the degree you enrolled in or other industries you may be switching from. To the contrary, many prospective employees often bring transferable skills and experiences that benefit their analytics careers. Day-to-day skills can be learned on the job but demonstrating a dedication to learning and picking up initial proficiencies will be attractive to prospective employers.

That being said, analytics is awash with job titles and varied, winding career paths. While there is often a significant degree of overlap activity between titles — and QuantumBlack offers plenty of freedom in charting your own career path — it’s useful to understand what’s expected of which discipline. Knowing where you’d like to start and the nuanced differences between roles will guide your initial career decisions and your choices when developing a data science portfolio.

Data Engineers (DE) are often tasked with building the infrastructure that powers an analytics solution, as well as finding, understanding and delivering data in a way that helps Data Scientists (DS) extract deep insights. DEs are typically then asked to take the insights that have been generated and deliver them to the end user, while automating the process and ensuring it runs robustly.

DE Sam Hiscox explains: “DEs aren’t just Structured Query Language (SQL) specialists. We’re better thought of as Software Engineers who have specialist skills in data. This allows us to design and build data-powered products that are resilient and scalable. Joining tables and understanding data structures are fundamental to our job, but writing readable, efficient and tested code is just as important.”

In contrast, DS Garazi Gomez de Segura Solay frames her role as “responsible for translating a business problem into an analytical problem, before building the models that create the solution. Data scientists often rely heavily on statistical and mathematical analysis. We rely on engineers to ingest, transform and join the data that feeds the analytical models we build.”

Building A Data Science Portfolio

A portfolio is ultimately an asset for the job interview — it provides the candidate with something to talk about and a way to demonstrate an initial skillset or proficiency. Building one from scratch may sound daunting and identifying that initial use case or a project can be difficult. However, it is useful to remember that data is fairly ubiquitous in most tech projects today.

“Any software project will involve a degree of data and I remember setting up my first database at 3am during a university hackathon,” explains DE Sam. “During my spare time I consumed plenty of Medium articles and YouTube videos that covered technology I was interested in, such as Docker, Airflow, Ml flow and others. I’d then build little demos using the technology I read about and tried to find opportunities in my previous job as a mechanical engineer to use my coding skills to automate data activities.”

DE Shawn Tan also attended AI-themed hackathons at university and also suggests gaining initial experience with startups, even if this is temporary and voluntary. “Even those positions I volunteered for on a short-term basis provided opportunities to build tech solutions such as web scrapers, chatbots and databases,” he says. “Grappling with new technology as you’re building it may sound challenging, but there’s plenty of existing documentation available online.”

Resources To Help Build Your Prelimiary Portfolio

Online platformslike Coursera, FreeCodeCamp, Udemy and Hackerrank were all used by our colleagues used when preparing for careers in data science. “Kaggle is also a fantastic resource,” explains Garazi, “as it provides cleaned data and a pre-configured environment, leaving you to concentrate on building the best model you can. Kaggle tends to be a popular platform for technically proficient practitioners to share their code on, so there are plenty of opportunities to learn from others.”

All three colleagues were clear that taking advantage of existing support is the easiest way to build an initial portfolio. “It’s important to remember that in analytics we almost never build software from scratch,” says Sam, “Often someone much smarter than me has thought deeply about the problem I’m facing, created a very elegant solution and then made it freely available as a Python package. I’d suggest that candidates get good at finding these assets, understanding the documentation around them and leveraging them. Building a product that solves a problem by cobbling together a selection of open source software is far more impressive than being able to implement a bubble sort in five different languages.”

Alongside an impressive portfolio, all three colleagues agreed that the most crucial asset a candidate can bring to a job is a visible eagerness to learn and improve. “I frequently hear some of the smartest people at QuantumBlack say ‘I don’t fully understand this, can you tell me more,’” says Shawn. “Whether technical knowledge or background on a fresh industry, the data practitioners I know are learning something new every day, so it’s not just accepted but expected that entry-level candidates won’t know everything. Arriving at an initial interview with a reasonable first attempt at a portfolio demonstrates a keenness to proactively learn, and that’s a trait that data employers value.”

If you’d like to discuss a career at QuantumBlack as well as other opportunities, please visit our careers page or alternatively reach out directly to a member of our team — we’re always eager to offer more information or advice.

Authored by: Garazi Gomez de Segura Solay, Senior Data Scientist, Sam Hiscox, Senior Data Engineer and Shawn Tan, Data Engineer, QuantumBlack