Hello, reader!

Quantum #662 had an open rate of 49% and a click rate of 12%.

The most clicked link was Erin Brokoich's AI Data Centre Reporting website.

Former data journalist and data analysis thought leader Gurman Bathia is running a very interesting hybrid conference (online and, in person, in Bengaluru) on 3-4 July. Take a look at VIZCHITRA 2026, it's a great mixture of workshops and talks. Read also this blog post where Gurman explains why this community exists.

'till next week,
Giuseppe

Topical

Bracketology: predict a path to World Cup victory

The Guardian's interactive World Cup simulator. It's very interactive. The article also explains how the tournament structure is designed to "maximise the spectacle" by preventing top teams from meeting too early.

Uruguay produces more famous footballers per capita than any country in the world

Interesting analysis of "Famous footballers in the global top 1,000 born between 1850 and 2015, per million inhabitants", with an interactive chart available. Full article here.

Stat Significant is a free weekly newsletter featuring data-centric essays about movies, music, TV, and more.

When do we stop finding new music? Which TV shows got their finale right, and which didn't? Which movies popularized (or tarnished) baby names?

Subscribe for free to find out!

Tools & Tutorials

SQL to ER Diagram

This is a free, open-source online tool that converts SQL schemas into interactive entity-relationship diagrams directly in the browser. You can paste CREATE TABLE statements and instantly visualise tables, columns, primary keys, foreign keys, and relationships. It supports multiple dialects.

kage

Kage is an open-source tool that creates offline, script-free copies of websites by rendering pages in headless Chrome, capturing the final DOM, and stripping all JavaScript.

visx

"visx is a collection of reusable low-level visualization components. visx combines the power of d3 to generate your visualization with the benefits of react for updating the DOM." It's by Airbnb.

Count Anything

This paper is about object counting research. While acknowledging the fragmentation in object counting research, it proposes a unified, text-guided approach (Count Anything) that works across diverse visual domains.

Quarkdown

"Markdown with superpowers". Quarkdown is a free, open-source tool that combines the simplicity of Markdown with the power of LaTeX for document creation.

How to Build a Simple, Bulletproof Data Pipeline

This article illustrates a simple, robust data pipeline design over premature optimisation and unnecessary complexity. "In most organizations, a lot of value can be delivered with a setup that avoids high costs and unnecessary complexity. The most common scenario is not real-time streaming or exotic architectures. It is daily extraction from one or more transactional systems backed by a relational database. Despite how common this scenario is, many pipelines fail for reasons that have little to do with technology. They fail because of fragile assumptions, premature optimization, or the introduction of complexity before it is actually needed. In this article, I want to walk through a simple but realistic example and show how a few design decisions, even when they look basic, can make a meaningful difference in robustness, operability, and long-term maintainability.".

British Columbia, Time Zones, and Postgres

On March 8, 2026, British Columbia permanently shifted to Pacific Daylight Time (UTC-7), eliminating the fall-back to UTC-8. This change exposes a critical flaw in storing future appointments using PostgreSQL's timestamptz columns, which store UTC values and convert to local time using timezone rules at query time. If timezone rules change between storage and retrieval, the displayed local time differs from user intent, potentially making November-March appointments appear an hour off. The article advocates for a "dual-column pattern" for future local events: storing both local timestamp and timezone name, plus a calculated UTC timestamp, preserving user intent while enabling efficient querying.

SVGs and PDFs can both be interactive

"SVG and PDF both have a rarely-used interactive layer hiding in their specs. SVG in particular can almost host small, fully contained apps. Did you know this? I didn't until somewhat recently, and I haven't found a cohesive treatment of the topic out there. So, here it is."
SVGs can function as self-contained mini-applications, manipulating DOM elements through scripting—a paradigm familiar from web development. PDFs also support JavaScript through Adobe's specified APIs, though primarily designed for forms rather than general interactivity.

lore

Lore is EpicGames' open source version control system, designed for unprecedented scalability of both data and teams, and optimised for projects combining code with large binary assets, therefore particularly suitable for game development.

Image compression.

Part of the Making Software series, this article shows how digital images are created by "exploiting the quirks of human vision to make images smaller, without looking awful."

Can gzip be a language model?

This is pretty nerdy. This article by Nathan Barry explores whether the gzip compression algorithm can function as a language model without neural networks or learned parameters, leveraging the compression-prediction equivalence principle: "every prediction model is inherently a compressor, and all compression algorithms are prediction models".

The verifiable data agent

"Probably is a secure, local app for accurate data analysis using natural language:
- Explores and reports quickly over billions of rows
- Determines verifiable answers to any data question, just ask
- Connects to any data source: local files or a massive data warehouse
- Builds understanding of your business context with every question"

The NULL in your NOT IN

"A NOT IN query can return the wrong answer without telling you. It is valid SQL, it runs without an error, and it hands back a perfectly well-formed result set that happens to be empty when it should not be. No warning, no hint, nothing in the logs: just zero rows where you expected hundreds, and a database that considers it correct.
Almost always the cause is a single NULL sitting somewhere you forgot to look, combined with two keywords you have typed a thousand times: NOT IN. None of it is a Postgres bug. This is exactly what the SQL standard mandates, implemented faithfully. That is precisely what makes it so easy to walk into, and why the planner could not safely optimize around it for the better part of Postgres's history. It comes down to one if statement in the parser."

datapitfalls

Datapitfalls is an open-source tool, powered by Claude AI, that detects common mistakes throughout the entire data reasoning chain, from initial question formulation through final chart presentation. Unlike typical chart linters that only check visual elements, this tool examines the full workflow: question formulation, data collection, transformation, analysis, visualisation, interpretation, and communication.

Data Thinking

A Catalog Is All You Need

Anders Swanson: "Way back in 2018 it took me a day or two of banging my head against my keyboard to connect two on-premise SQL Servers with external tables.1 “How could it be hard?” I thought. As I learned more in the intervening years, I came to appreciate the work (and luck) required to pull off the user experience I dreamed of. Today, I’m here to say it largely just works 2 — almost boringly so. Exactly how it should be.
So today I want to show you a “simple” dbt demo that’s been years in the making. The big shift it points to: the catalog is becoming the unit of interoperability — not for one query engine, but for all of them."

Repeat yourself

Evelina Parrou explores how strategic repetition enhances data communication and persuasion. Whether this is about repeating the legend as labels, or message redundancies, it is useful to clarifying the information.

Data manual

The UK government's Data Manual, hosted on data.gov.uk, and part of the work by the National Data Library team. It functions as a comprehensive resource hub for government data practitioners, covering essential topics across nine main categories.

Dataviz, Data Analysis, & Interactive

What are left-handed people like? 👈

Leo Benedictus analyses data about right- and left-handedness.

A1 Collision Density, 2000-2024

Benjamin Niedermann uses DfT's STATS19 dataset to visualise accidents on the A1. See also his LinkedIn post. The viz uses the VisQuill library.

Global Manufacturing Statistics by Country

Hell of a Partner, a B2B directory, tracks 50,751 verified manufacturer and distributor profiles across 217 countries and 93 product categories. They release an open, CC BY 4.0 index of 32,000+ manufacturers and 2,400+ distributors across 199 markets, built from national registries and trade associations, with every website checked live before listing. The author wrote noting that the share of companies with certification X is mostly a reporting artefact (how fully each source recorded the field), not a country ranking, which is a useful caution for anyone aggregating supplier data. Fundamentally, the within-market shares are real (around 70 to 83 percent of indexed food-specialty makers list HACCP across ES/IT/US/IN); the cross-country comparisons are not.

Europe's housing shortages are even worse than America's

"Why does nobody there talk about zoning?", asks Works in Progress, arguing that European housing policy debates should adopt American perspectives on zoning reform. While European policy discussions typically focus on rent controls and public housing, the author contends that Europe faces worse housing shortages than America.

I found a seashell in the middle of the desert

"To my amazement, I found a fully solid rock that eerily resembles a seashell at the base of a cliff in the Alghat desert, Saudi Arabia. I didn't know what to make of it at first, it had the swirls and shape of a seashell but was fully a rock, more importantly, it shouldn't be here; the nearest coastline is Dammam's, 500 km away." What follows is a morphological analysis, using a dataset of over 7,000 shell species over almost 60,000 images, applying PCA and other algorithms, to find out what type of shell this is.

Current Rothko

An interactive website matching Mark Rothko paintings to the viewer's current location, weather, and mood.

Can you see three trees?

Amanda Shendruk examines the 3-30-300 rule, a popular urban planning standard proposed by Cecil Konijnendijk for measuring equitable access to nature in cities. The rule states that every home, school, and office should have a view of at least three trees, be in a neighbourhood with 30% tree cover, and be within 300 meters of a park.

AI

The Epoch Brief - June 12, 2026

This newsletter from Epoch AI highlights several developments in AI capabilities and infrastructure as of June 2026, with a very interesting analysis of the impact of LLMs like Mythos on Cyber Vulnerabilities reporting.

Introducing the Open Knowledge Format

Google Cloud is introducing the Open Knowledge Format (OKF), an open specification designed to improve how AI systems access and share organisational knowledge. OKF formalises the "LLM-wiki" pattern into a portable, interoperable standard using markdown files with YAML front matter.

Three Ways to Think About AI and Jobs - The Atlantic

Rogé Karma for The Atlantic "Whether automation will make human workers obsolete depends on more than just how smart the AI is." Using the example of radiology, where despite predictions that AI would replace radiologists within five years, the profession has grown in numeric and salary terms, the article presents three key question for assessing AI's threat to any profession: is it a job that has a "strong bundle" of interconnected tasks are interconnected? Will increased efficiency through AI create higher demand, as it happened with ATMs and bank tellers initially? Does AI replace workers' least expert skills (enhancing their value) or most expert skills (commodifying their labor)?

What can 500 years of journalism teach developers about AI trustworthiness? - Stack Overflow

"AI reliability issues stem from three separate architectural challenges that keep getting lumped into the same category. Prompt engineering alone can't fix them. But the sourcing and verification frameworks media organizations have used for centuries translate into clear engineering solutions developers can implement today."

🔮 Is AI immune to groupthink?

Azeem Azhar: "I am a fan of AI councils. As many readers will know (from previous posts), I use AI expert panels every day to deliberate on hard questions, challenge my thinking, and do research. These AI councils, or panels, start from an assumption that a group can produce better work than a single agent. My friend and one of the sharpest AI practitioners out there, Rohit Krishnan, tested this hypothesis."

AMA – Ask Me Anything! Submit a question via this anonymous Google form. I'll select a few every 4-5 weeks and answer them on here :-) Don't be shy!

The Quantum of Sollazzo grove now has 40 trees. It helps managing this newsletter's carbon footprint. Check it out at Trees for Life.

'till next week,
Giuseppe @puntofisso.bsky.social