⚡️Trendbreak #17⚡️

Yeia! 🇬🇷

For those of you freshly graduated, it might have been a bit unnerving hearing your more seasoned colleagues throw around terms like OLAP cubes, slice and drill-down 😱. This post from the Analytics Engineer Club does a fantastic job at breaking down these definitions (kudos to my colleague Dejan for pointing it out!). They are actually simple concepts, although often poorly or not at all explained (particularly as they have not been widely taught for several years now). The idea is to pre-aggregate data in order to speed up analytical queries. This was particularly helpful a few years ago when memory and computation resources were more limited, but it's less relevant today, unless you're working at petabyte scale.

Next up, some exciting news from the intersection of machine learning and pharmacology 🤖🧪: ChemicalX is a recent collaboration between AstraZeneca, Vanderbilt University, and Harvard Medical School (thanks to my colleagues Jérôme and Nabila for the tip!). It's an open-source Python library containing tools to tackle the issue of drug pair scoring, in other words, answering the question "what happens when two drugs are administered at the same time?". Interactions can indeed be positive (we then talk about synergy) or negative (polypharmaceutical side effects). Several models encapsulating different state-of-the-art approaches are assembled under a common PyTorch-based API, one of the leading deep learning frameworks.

Lastly, if you're still grappling with the Pandas 🐼 syntax, the primary tool for exploratory data analysis in Python 🐍, you should take a peek at Bamboolib 🎋. It's a library designed for interactive use within a Jupyter notebook: you manipulate data using a small graphical interface, and Bamboolib writes the corresponding code (this will sound familiar to anyone who's ever recorded macros in Excel!). This tutorial explains how to implement the main data transformations (column operations, row filtering, merges, group by operations, and so on).

As always, happy reading and have a fantastic week! 📚

By @Clément Chastagnol in
Tags : #Trendbreak,