The world’s leading publication for data science, AI, and ML professionals.

Stats for Data Science, from the Ground Up

Our weekly selection of must-read Editors' Picks and original features

Photo by Alisa Anton on Unsplash
Photo by Alisa Anton on Unsplash

Data scientists love debating which skills are essential for success in the field. It makes sense: in a rapidly changing ecosystem that adopts new and powerful technologies all the time, job requirements and toolkits never stop evolving.

Statistics seem to be one major outlier, though. Data professionals of all stripes seem to agree that a solid foundation in stats and math will make your life easier regardless of your role, and can open up opportunities that would otherwise remain beyond reach.

To help you on your learning journey, we’re sharing a few of our favorite recent posts that focus on statistics for Data Science and machine learning. They go from the basics all the way to more specialized use cases, but they’re all accessible, beginner-friendly, and emphasize practical applications over lofty theory. Let’s dive in!

  • Stats novice? Not for long! If you’re tackling stats for the first time in your professional life—and especially if your memories of high school math inspire more dread than joy—you’re bound to appreciate Chi Nguyen‘s simple explanations of basic concepts.
  • A structured approach to learning statistics. Looking for a thorough, step-by-step resource for learning stats? Adrienne Kline recently launched an excellent Statistics Bootcamp that unpacks the math behind all the data science libraries practitioners use daily. (If you’ve already discovered the first installment, linked above, parts two and three are already out!)
  • Making sense of occasionally confusing terms. For his debut TDS article, Ajay Halthor shared a lucid explanation of likelihood, and focused on the role it plays in machine learning, as well as its sometimes hard-to-grasp connection to probability, an equally crucial concept.
  • Putting your statistical know-how to good use. There’s always a gap between theoretical knowledge and its effective application. Mintao Wei‘s recent contribution does a great job bridging it, as it walks us through the process of selecting the right statistical tests for a range of A/B testing metrics.
  • The inner workings of a powerful algorithm, explained. The bootstrap, says Christian Leschinski, "is an algorithm that allows you to determine the distribution of a test statistic without doing any theory." It’s also one that’s been "widely overlooked." Harnessing his deep knowledge as a statistician, Christian guides us through the magic behind the boostrap, and shows how it can help practitioners in their analyses.
  • Why it’s crucial to connect statistics to business outcomes. Cassie Kozyrkov identifies the challenges data professionals face when they bring their stats and math knowledge to work projects, and stresses the importance of data budgeting, a topic college classes rarely cover. (If you’d like to read more of Cassie’s insights—and you should!—don’t miss our brand-new Q&A with her, which touches on data career paths, the value data analysts bring to companies, and much more.)

All stats-ed out, are we? We hope not, but just in case—here are some non-statistics-related reading recommendations we think you’ll enjoy.


Your support means so much to us – thank you for reading our authors’ work; a special shoutout goes to all of you who’ve recently become Medium members.

Until the next Variable,

TDS Editors


Related Articles