How to Build Successful Data Teams
The process of becoming a data scientist revolves around growing your individual skill set and knowledge base. Yet if you gather a group of sharp, highly skilled data pros and expect smooth workflows and intuitive collaboration to just happen, you’ll likely be unpleasantly surprised. Building a team takes work—a lot of it.
This week, we’ve selected our best recent articles on the elements that make data teams thrive. Some of our picks focus on collective frameworks, while others highlight the necessary skills for individual contributors to flourish within the larger group. You can choose your own adventure depending on the current challenges your team is facing, or just… read them all. They’re worth your time.
- Bringing the right people together. A collection of experienced, skilled professionals won’t become more than the sum of its parts without a plan. As Angelica Lo Duca emphasizes, putting together a team requires hiring people into roles that align with their strengths, hitting the right balance between specialists and generalists, and assigning individuals to the tasks and projects where they can make a real difference.
- How to measure success. KPIs—key performance indicators—have become a widely used tool for measuring progress towards an organization’s shared goals. Barr Moses argues that for data teams, defining and executing on the right KPIs is even more crucial, and sometimes more difficult. Her article offers several insights on how to create effective KPIs for data teams.
- The ability to move between different data types is essential. “At the end of the day, data consumption is the final deliverable for data teams,” says Luis Velasco. He goes on to unpack the common challenges teams face when building a data architecture that other stakeholders can trust, and introduces the idea of a polyglot framework to address some of the thorny aspects of the data mesh paradigm.
- What might first principles for data teams look like? Adam Sroka’s four rules for being an effective data scientist are geared towards each team member’s personal standards of execution. They apply equally well on the team level, though, with a strong preference for usefulness over complexity and for simplicity over novelty, among other key ideas.
- Adding value by framing questions the right way. We can accept that there’s no such thing as a stupid question, but that doesn’t mean that all questions are created equal. “How” and “what” have their place, but for Genevieve Hayes, PhD, “If you want to maximize the value you bring as a data scientist, you need to start asking why.” This will have a positive effect not just on the perception of your own work, but on that of other data practitioners in your company, too.
- How to center the needs of your internal clients. Every data team is located somewhere along the spectrum between service-oriented and product-focused. Maeda Hanafi’s detailed walkthrough of a recent project—the redesign of a clunky visual debugging tool—shows how teams can be both when they create products that empower other partners and stakeholders.
- Tell a compelling story with your team’s work. All your analyses, insights, and recommendations won’t make the biggest possible impact if your team doesn’t communicate them well to your colleagues—especially if they’re less data-literate than you. To polish your presentation skills, we offer a double bill of great reads: Brian Roepke’s eight tips for creating a compelling presentation, and Eirik Berge, PhD’s five essential presentation tips for data professionals. (That’s a total of 13 useful tips, for those of you keeping score.)
Looking for a few more reading recommendations on other topics? Happy to oblige:
- We just launched our new Writer’s Workshop series with a thorough guide on finding an audience for your data science, ML, or other technical articles. Are we objective? No. Is it really, really good? Yes.
- In case you missed it, we recently chatted with healthcare data analyst Rashi Desai about finding success in your chosen data career path.
- If you’d like to take your Matplotlib skills to the next level, here’s an advanced tutorial by Bex T.
- Inclusive design in your data visualizations can make a difference—here’s Leonie Monigatti’s hands-on primer on creating color blind-friendly charts and plots.
If you’d like to support our authors’ work, consider becoming a Medium member; it also makes the entire TDS archive (as well as every other paywalled Medium post) accessible to you.
Until the next Variable,
TDS Editors
