📚 Let’s Expand Our Data Science Horizons

Welcome to the 100th edition of the Variable!

We usually devote each edition of our newsletter to a specific topic or set of questions; to mark this special occasion, though, we opted for a different approach. Every week, we leave fantastic TDS posts out of the Variable simply because they don’t quite fit within our theme; this week, We’re presenting the best of our recent Editors’ Picks that we haven’t yet highlighted.

In case you’re worried we might be leading you down a dangerous path to chaos and curatorial recklessness, have no fear! Each of our recommended reads invites you to explore new directions, approaches, or tools in a fresh, engaging way—and we believe that’s as strong a thematic link as any. Enjoy!

• Prediction Performance Drift: The Other Side of the Coin. Why does the accuracy of a model’s predictions deteriorate over time? Valeria Fonseca Diaz’s clear overview presents a helpful typology of four kinds of drift, including their causes and potential solutions.
• Will a Training Break Ruin My Marathon? barrysmyth’s latest contribution sits at the (fascinating) intersection of data analysis, sports, and self-tracking, and examines the effects of disruptions on marathoners’ performance.
• We Should Have Seen ChatGPT Coming. In a memorable debut post, Samantha Hodder blends pragmatism and wistfulness to assess ChatGPT’s still-unfolding impact—and to propose a more levelheaded mindset going forward, one that would allow us to “make it work for us and with us, not against us.”

• Survival Analysis: Predict Time-To-Event With Machine Learning. What’s better than predicting the probability that an event will take place? For Lina Faik, it’s knowing how much time remains before said event; Lina’s accessible explainer post walks us through the process of making such predictions in the context of customer churn.
• Two First Serves: Analyzing ATP Service Data from 2000–2020. Major team sports like basketball and football (both the universally played variety and the American one) often seem to dominate discussions of sports analytics. Sean Holland helps broaden the conversation with a thoughtful, data-backed look at tennis players’ key shot: the serve.
• Equal-Size Spectral Clustering. Who doesn’t like a well-illustrated tutorial on a perennially important topic? Carmen Adriana Martinez Barbosa’s new guide introduces us to the equal-sized spectral clustering algorithm, which produces—you guessed it!—more balanced clusters in terms of the number of points they contain; that, in turn, can prove essential in a number of real-world use cases.
• A Day in the Life of a Chief Data Scientist. The career paths of data professionals can take many shapes; if a management track sounds interesting to you and you’d like to get a better sense of what your work might look like down the line, Ray (one half of writing duo Leah Berg and Ray McLendon) shares the routines and practices that shape his experience as Chief Data Scientist.
• How to Make an AI Image-Editing Chatbot. To end on a hands-on, creator-friendly note, don’t miss this quick walkthrough by Sophia Yang and Philipp Rudiger, who show us how to create an app that can edit images using chatbot prompts.
• Bonus pick: “The main driver behind my writing has always been learning.” We were thrilled to publish our first Author Spotlight of 2023 this week—a lively conversation featuring Matteo Courthoud, who generously shared insights about his career path, writing journey, and ongoing interest in causal inference.

Thank you for joining us on this 100th edition of the Variable—here’s to 100 more!

If you enjoyed the work we highlighted this week, we hope you consider becoming a Medium member — it’s the most direct and effective way to support our publication.

Until the next Variable,

TDS Editors

Let’s Expand Our Data Science Horizons was originally published in Towards Data Science on Medium, where people are continuing the conversation by highlighting and responding to this story.