2021#01 Readings4 minutes read | 690 words by Ruben Berenguel
As promised, the numbering of these posts is now year-indexed.
I have started the year strong, writing a post about logging in Spark in Databricks, and writing a web-based version of Guillotine. See them below.
📯 Configuring log4j properties in Databricks (and EMR)
First post of the year, a particularly problematic area. Logging is hard in Spark: getting Spark to log, where it puts the logs, configuring the logs, reducing the verbosity… So many “classic” questions. In this, I tell you how to configure logging via a
log4j.properties file in Databricks and EMR. With it, you can configure external logging aggregation services, like ElasticSearch or Graylog.
Two weeks ago I finished Guillotine, a standalone, cross-platform Electron app that lets you have a circular (or rounded square) shaped window with your webcam feed, for recording your screen in style for online presentations. It had some drawbacks, particularly it would not work in full-screen presentations because a window can’t float over a full screen.
To solve it, I wrote this one. It shares a fair bit of the code, but it creates a floating and dragable div with your camera, so you could use it in full screen for web based presentations (like Google Sheets, although you need to be a bit hacky, for self-hosted reveal.js it works wonders, there is a demo in the README).
🍿 Data Science with Rust - Arrow, DataFusion, and Ballista
I didn’t know Ballista and DataFusion were so close to being usable.
Why cohorts is a privacy detour the media industry should avoid
This feels written by someone with an agenda.
You’re Allowed To Make Your Own Tools
I can wholeheartedly agree.
Learning to Play the Chaos Game
The Chaos Game is an approximation method for IFS (iterative function systems), and this looks at the reverse problem: getting an IFS out of a particular image. Interesting to see! And wondering about some applications to stuff I used to do 🤔🤔🤔
🐦 A thread on how creating generative art takes time and knowledge
As an occasional generative artist, this is spot on. Collecting techniques and eventually mixing them takes time.
Things I Learnt in 2020
A surprisingly large amount of miscellanea.
How we built the GitHub globe
You can check a previous edition of the readings to see a similar post from Stripe’s own globe.
A year of tips
From Josh Comeau’s newsletter, it has several “tips” with HTML or CSS things you may not know. I have taken 4 or 5 to use ASAP.
Can 4 Seconds of Exercise Make a Difference?
Dr. Coyle has equity in the company that manufactures the bicycles, but says this monetary involvement does not affect research results from his lab.
It’s… a risky stance for a researcher, but hey, I may be able to handle 8 seconds (the estimate for what you’d need in a normal stationary).
The Lasting Lessons of John Conway’s Game of Life
Commentary from researchers and creatives about what they think about Life
Reverse Engineering the source code of the BioNTech/Pfizer SARS-CoV-2 Vaccine
I didn’t know much about mRNA techniques (my years of interest in this area were some time ago, when I was thinking of studying biotechnology), and this has been an enlightening post.