What is in here
- The Long Bet – where and what to study when AI keeps changing which skills stay scarce: an ongoing series, each part answering one piece of that question against data collected for it.
- 100 universities from the Shanghai Ranking top 100, across 19 countries, and 31 STEM programmes – 1,200+ university–programme pairs checked one by one and normalised to 31 canonical labels before any analysis ran.
- The data is downloadable: 2,303 rows across 23 editions of the ranking, 2003–2025, as CSV and xlsx.
- Every obvious answer was tested against the data, and most turned out wrong – the series documents where the received wisdom failed and why.
A multi-criteria optimisation problem – university access, a future-proof programme, flexibility, cost – and why obvious answers don't survive the data.
Article 1 · Jul 2026 · Original research
Read the full article →
Most university comparisons are built on a mistake. They compare labels, not substance. Here is how I avoided that – and what it took to build a dataset that actually holds up.
Article 2 · Jul 2026 · Original research
Read the full article →
I built the coverage matrix expecting to confirm a hunch. It didn't: CS at 95 universities, AI at 12, France above Imperial. What the data showed.
Article 3 · Jul 2026 · Original research
Read the full article →
Why the A-level combination top universities have always wanted hasn't changed – and why most students don't pick it.
Article 4 · Jul 2026 · Original research
Read the full article →
The scarce skill in data and computing has shifted seven times since the 1880s. The foundation underneath it has never moved.
Article 5 · Jul 2026 · Original research
Read the full article →
Every university in ranks 1–100 of the Academic Ranking of World Universities, all 23 editions, as one table and one CSV – the numbers behind this series' country comparisons.
Dataset · Sep 2026 · Original research
Open the dataset →
Which STEM programmes the world's top 100 universities actually teach: 1,258 university-programme pairs normalised to 31 canonical names, as CSV.
Dataset · Sep 2026 · Original research
Open the dataset →
What STEM degrees ask for at A-level: 23 programmes read off live admissions pages, Mathematics named by 22 of them, as CSV.
Dataset · Sep 2026 · Original research
Open the dataset →
The Top 100 Is a Moving Target
Twenty-three editions of ARWU, year by year: the US from 52 top-100 seats to 37, China from zero to thirteen since 2016, one French merger from #37 to #14. The ranking behind this shortlist has never been still.
Planned · Original research
Coming soon
China: The Second Look
China was left unresolved – cheap, fast-climbing, absent from international shortlists. How admission works for international applicants, what the top-tier universities teach, and why families don't see them.
Planned · Original research
Coming soon
The Full-Ride Playbook
Sticker price eliminated the US on cost alone – but sticker price and actual price are not the same number. What need-blind admission and full-need aid require to bring the real cost close to zero.
Planned · Original research
Coming soon