514: quantum of sollazzo
#514: quantum of sollazzo – 25 April 2023
The data newsletter by @puntofisso.
Hello, regular readers and welcome new ones :) This is Quantum of Sollazzo, the newsletter about all things data. I am Giuseppe Sollazzo, or @puntofisso. I've been sending this newsletter since 2012 to be a summary of all the articles with or about data that captured my attention over the previous week. The newsletter is and will always (well, for as long as I can keep going!) be free, but you're welcome to become a friend via the links below.
I type this newsletter from Buenos Aires airport, as I wait for my flight back to the UK after I came hear to give a keynote at csv,conf,v7 (yes, it's really the name of the conference, and it's wonderful). I gave a 40min talk on the 12 lessons I've learned working as a data wrangler, as a data advocate, and as a data leader in government. I found the conference engaging, populated with an outstanding community.
The conference, despite its name, is not just about data wrangling. So together with excellent practical sessions like one about Miller by its creator John Kerl, we also had talks about how data is helping transparency in Brazil by the OKFN's Fernanda Campagnucci (extra points if you get the pun) and in Kenya by Code4Africa's Tricia Govindasamy, and we learned about how data can help scrutinise the work (and misconduct) of law enforcement in Georgia with a talk by LLEAD's Rajiv Sinclair and Ayyub Ibrahim.
Expect videos soon, and when they are ready I will add a full round-up section for the best talks. But for now you get this picture of me opening my keynote and a link to the slides and references.
Mostly thanks to Lewis Bush's incredible work and tutorials, I got myself practicing the art of cynotyping... using Open Street Map to create the negatives. I'm really enjoying it. Give me ideas of nice areas to picture, and I'll execute them :-)
You should also totally check Lewis' recent book Depravity's Rainbow. As per Wikipedia, "His book Depravity's Rainbow (2023) is about early rocket development in Nazi Germany including the V-2 ballistic missile and the way that many engineers involved in these projects were recruited by Allied countries after the war and went on to play a major role in post-war rocket development including at NASA during the Apollo project. The book predominantly focuses on Wernher von Braun."
The most clicked link last week was How to use AI to do practical stuff: A new guide.
'till next week,
Giuseppe @puntofisso
|
Become a Friend of Quantum of Sollazzo from $1/month → If you enjoy this newsletter, you can support it by becoming a GitHub Sponsor. Or you can Buy Me a Coffee. I'll send you an Open Data Rottweiler sticker. You're receiving this email because you subscribed to Quantum of Sollazzo, a weekly newsletter covering all things data, written by Giuseppe Sollazzo (@puntofisso). If you have a product or service to promote and want to support this newsletter, you can sponsor an issue. |
✨ Topical
Here's when the debt-limit crisis gets increasingly risky
"Highs and lows of U.S. cashflow will make for a bumpy ride, intensifying the partisan stalemate as the nation moves closer to defaulting on its $31.4 trillion in debt."
Yay, make Government budget an adventure...!
Yes, the movies are getting longer and longer
"The most successful movies of recent years are longer than at any other time in history. But are they better?"
Originally in Spanish and here automatically translated into English. It's part of this good round-up of 2022 dataviz.
(via Alex Selby-Boothroyd)
Do we use our right to vote?
"What counts as a high turnout? Do we take our democratic rights for granted when people in other countries don’t have that luxury? How do other countries compare?"
Including one of my favourite topics, voluntary vs mandatory voting.
Seismic Illumination
"Pacific Rim Earthquakes as the Echo of A Planetary Process". Aside from ArcGIS, it uses a bunch of open/free sources/.
Chronology of a Radicalization
"How Telegram became the most important Platform for Conspiracy Ideologies and Right-wing Extremism". By the nonprofit Center for Monitoring, Analysis, and Strategy (CeMAS), funded by the Alfred Landecker Foundation.
Slow Zones
"Paris is SLOWING DOWN to SPEED UP the recovery of city streets."
Controversial in the UK, so here's some useful data-driven look at it.
Groundwater Gold Rush
"Banks, pension funds and insurers have been turning California's scarce water into enormous profits, leaving people with less to drink."
Congress Today Is Older Than It's Ever Been
FiveThirtyEight looks at the age of Congress members.
🛠️📖 Tools & Tutorials
Midjourney AI Guide
A beginner's guide with prompts, resources, and hacks.
Colorcontrast.app
"Color contrast checker. Check and suggests colors to meet the required WCAG or APCA contrast ratio."
Scraping without programming
A handy tutorial (in a Google Slides presentation) that focuses on how to use Google Sheets to harvest web data.
PandasGUI
"PandasGUI is a GUI for viewing, plotting and analyzing Pandas DataFrames."
Spinning Diagrams with CSS
An interesting use of animations in CSS, which is useful to represent simple projections of 3D objects from all their sides. I suspect this could be made even better by allowing the user to manually rotate the object.
EQUALITY OF ODDS
"A Visual Introduction to Measuring and Mitigating Bias in Machine Learning".
Another great tutorial by the MLU-Explain folks.
d′etectable
"An Explorable Explanation of Signal Detection Theory"
Awesome Twitter Algo
An annotaded look at the Twitter Algorithm and the way it creates its recommendations.
🤯 Data thinking
The NYPD Was Systematically Ticketing Legally Parked Cars for Millions of Dollars a Year- Open Data Just Put an End to It
Old (from 2016) but still very important, together with this other article about how Open Data Revealed $791 Million Error in Newly Adopted NYC Budget.
(via John Kerl)
How intrepid Victorian surveyors mapped the length and breadth of Britain
"Today’s country walkers owe much to the theodolite-lugging cartographers of the early Ordnance Survey."
📈Dataviz, Data Analysis, & Interactive
Global Subway Spectrum
An exploration of colors used for lines in every rapid transit system.
Inteligencia artificial para mapear parques solares en Argentina
Argentinian media outlet La Nacion has been working on using AI to automatically map PV panels farm, here in the original Spanish and here automatically translated into English.
(via Massimo Conte)
UMAP
"uMap lets you create maps with OpenStreetMap layers in a minute and embed them in your site.". For example, below from this tweet, a visualization of urbanistic elements of Venice.
(via Alessandro Luollin)
The Hidden Ways We Really Work Together
"Microsoft researchers used machine learning to find out how collaboration happens—beyond the org chart."
(via Massimo Conte)
Visualizing Max Temperature Trends in the USA
You'll judge if this is the most accurate way to make inference from the data, but the resulting animated visualization, here shared with its source code, it's really good at showing the type of variation that climate data represents.
Pokémon Explorer
"Explore 9 generations of Pokémon. Filter by type and stats, and scout any Pokémon by using the interactive visualization."
Detecting heart murmurs from time series data in R
"Time series analysis can uncover hidden structures in data collected over time. In this blog post, I'll use {tsfeatures} to extract time series features and {tidymodels} to predict which sound recordings of heartbeats contain heart murmurs, using those time series features."
🤖 AI
AI and the American Smile
"How AI misrepresents culture through a facial expression."
Interestingly, this is something I noticed when I worked at my various instances of average faces: some country average are more smiley than others.
Column: Afraid of AI? The startups selling it want you to be
I'm not sure that this take is right, but as much as I don't agree with dismissing the dangers of AI misuses, I also agree that some companies working in this space are all too willing to depict themselves and, by extension, AI, as evil genius.
Teaching ChatGPT to Speak my Son’s Invented Language
The things that ChatGPT is (claiming to be) able to do are increasingly interesting.
What does this say?
Oh, this is a brilliant find. Ask ChatGPT to create some ASCII art, and it can get it spectacularly wrong.
91% of ML Models Degrade in Time
Usual disclaimer: this is not an impartial article, as it's from the perspective of a tool that claims to monitor exactly the problem of model decay. But it's an open source tool (which is great for this sort of thing!) and the article offers some good thinking and references about the problem.
The picture below is from this paper on Nature, which the article references heavily.
90% of My Skills Are Now Worth $0
"...but the other 10% are worth 1000x".
I've often – and, recently, increasingly – been thinking about the themese in this article.
"I can list skills that have proven valuable to me & clients in the past. Many of them are now replicable to large degree. Others, like baking, not so much (but also rarely valuable in a consulting context)."
Large Language Models Are Human-Level Prompt Engineers
Prompt Engineers are the newest and hottest kid of the block. According to this pre-print paper, they are top of the league of likelihood of disappearing, as... Large Language Models increasingly make better prompt engineers.
(via Paolo Filipe Bertoldi)
quantum of sollazzo is supported by ProofRed's excellent proofreading. If you need high-quality copy editing or proofreading, head to http://proofred.co.uk. Oh, they also make really good explainer videos.
Supporters* casperdcl and iterative.ai Fay Simcock Naomi Penfold
[*] this is for all $5+/months Github sponsors. If you are one of those and don't appear here, please e-mail me