A Government Liquor Board Lab-Tests Every Bottle It Sells. Almost Nobody Uses the Data.
The best datasets aren't hidden. They're published in plain sight by bureaucracies that have no idea anyone might want them — collected for compliance, exposed out of habit, and abandoned one page at a time. Here's my favorite example. The LCBO — Ontario's government liquor monopoly, and one of the largest single buyers of beverage alcohol in the world — runs a quality-assurance lab. Products it…
The Ontario government-run liquor monopoly, the LCBO, conducts lab tests on every bottle of alcohol it sells, including sugar content in grams per litre for over thirty thousand products. However, the LCBO website lacks any interface to utilize this valuable data, leaving diabetics, keto dieters, and others unable to easily access this important information.
Winesugarcontent.com, built by the author, fills this gap by providing a searchable and sortable database of lab-verified sugar content, along with additional filters like price, category, country, grape, and sweetness level. The site's data pipeline involves ingesting the LCBO search API, cleaning and deduplicating the data, and generating static pages using Astro and PostgreSQL.
The project's key principles include never allowing language models to invent numbers, trusting lab measurements over marketing descriptors, and recognizing the interface as the product itself. The same data publication pattern can be replicated with other government purchasing monopolies, such as Sweden's Systembolaget or Finland's Alko, providing valuable information to the public.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written; read the original for the full account.
