Beginner-Friendly Data Science Reads (That Advanced Practitioners Will Enjoy, Too) | by TDS Editors | Apr, 2023

  • Get familiar with a popular machine learning framework. If you’ve been tinkering with gradient-boosted algorithms, chances are high you’ve run into LightGBM. In case you could use some guidance on how to make the most of it, Leonie Monigatti’s introduction to the most essential LightGBM parameters is clear, actionable, and well illustrated.
  • It’s never too early to think about ML project design. Of all the (many) industry buzzwords that have come into circulation in the past few years, MLOps seems to have one of the longest shelf lives. Chayma Zatout’s deep dive is a good starting point if you’re not sure how this concept might relate to your day-to-day workflows, and how to apply its principles to your current projects.
  • Ease your way into building a solid pipeline. Apache Airflow might be a common tool for data engineering teams, but as Aashish Nair points out, its ubiquity doesn’t make its terminology, features, and quirks any less daunting. To help, Aashish presents a Python-based demo that walks readers through the process of creating a simple Airflow pipeline.
Photo by Mel Poole on Unsplash
  • Neural networks from the ground up. It’s all but impossible to understand the major strides we’re seeing in AI research without a firm grasp of neural networks. Dr. Roi Yehoshua’s overview of perceptrons—“one of the earliest computational models of neural networks”—is a gentle entryway into the topic, and covers the basic concepts before moving to a Python implementation.
  • Streamline your learning process with better notes. Regardless of the data science topic you’d like to focus on in coming weeks, a good note-taking practice can make a real difference. Madison Hunter’s new post presents a six-step roadmap to more effective studying and better retention.
  • Get up to speed with an up-and-coming Python library. If you learned how to work with DataFrames using Pandas—a likely scenario for many data scientists!—you may or may not be happy to know that a new library, Polars, has been gaining a lot of traction in recent months thanks to its high-speed performance. David Hundley’s latest post is geared towards Pandas-trained people who’d like to explore Polars’ benefits.

Source link

Leave a Comment