Summer Reads for Data Scientists

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

Jul 7, 2022

4 min read

So-called "beach reads" often get a bad rap: lightweight. Substance-free. Fluff.

We beg to differ. The highlights we selected this week are fun; many of them focus on passion projects or dive deep into quirky, less-common topics. They’re perfect for passing the time on your way to the nearest lake, campsite, or jazz festival. (Apologies: TDS is a Canadian publication, and so is our concept of summer.)

These articles are also thoughtful and enlightening, however, and a great match for those of you who plan to learn new skills or discover off-the-beaten-path areas of data science this summer. As for our Southern Hemisphere friends: you’ll love these too; just skip the flip-flops and opt for a good blanket and a nice cup of tea. Let’s begin!

Baseball is still a perfect sport for data scientists. For millions in North America, Japan, and beyond, summer is synonymous with baseball. Even if your knowledge of the game is rudimentary, though, you’ll want to read Momin Mehmood Butt‘s analysis, which looks at batting stats and how their significance has changed after the rise of Moneyball-style sabermetrics.
Learn about stochastic processes—the right way. If we choose to tackle a complex concept while frisbees (and mosquitos) fly over our heads, it better be with a clear, well-structured explainer. That’s precisely what you’ll get with Xichu Zhang‘s patient introduction to Itô’s lemma, an important tool in stochastic processes and financial mathematics.
‘Tis the season… for writing! If you have some free time this summer, perhaps you can devote a chunk of it to sharing your knowledge with the rest of the world. Need inspiration? Look no further than Xiaoxu Gao‘s recent post about her technical-writing journey—it includes both practical tips and high-level insights that might just spark your own creativity.

A parking conundrum, explained. We have a soft spot for articles that examine small details in our daily lives through a data-science lens. A great addition to this genre is Ralph Asher‘s latest post: it takes a strange pattern in a gym parking lot as its point of departure, and goes on to explore discrete-event simulation in great depth.
Traveling salesman problem: the German edition. Himalaya Bir Shrestha lives in Germany, and uses the country as the backdrop for solving one of the most famous optimization problems out there: how would you travel between all 16 state capitals, passing through each one exactly once?
Yes, data cleaning can be rewarding. Just like with physical spaces, tidying up a massive (and messy) dataset doesn’t often spark joy. Luckily, Ang Li-Lian is here to the rescue. Her reader-friendly, illustrated guide will inspire you to build your own streamlined data-cleaning pipeline.
Everybody loves a good-looking schema. Data architectures can be complex and hard to grasp. Instead of a convoluted, "now this happens, then that happens" verbal run-through, consider using a visual schema. Benoit Pimpaud provides all the design tips and tricks you’ll need, and you can get started with nothing more than a pen and a piece of paper.
Take your first steps with design patterns. There are many good reasons to expand your knowledge of design patterns—from better communication with dev teams to a deeper understanding of the libraries and packages you use as a data scientist. Semi Koen‘s primer is a perfect place from which to build up your foundation in this area.
It’s never a bad time to ensure you’re ready for your next opportunity. Job hunting from the beach might not be for everyone, but what if the buzziest startup in town just posted the perfect role for your skill set? Ana Isabel Casado recently shared a helpful resource about preparing a CV that will boost your visibility during the screening process.
Explore the fascinating world of fractal music. We might still be a few years (or decades?) away from an AI composing the next song of the summer, but that hasn’t stopped Harlan Brothers from trying to "take a synthetic dataset of musical characteristics with a known distribution and algorithmically turn it into music," a process he recounts in his debut TDS post.