Fields Medal to Chinese Mathematicians: The Data Science Behind Harvard\'s 1,184 Oxalate Measurements
July 24, 2026 — Chinese mathematicians Deng Yu and Wang Hong won the Fields Medal yesterday in Philadelphia — the first Chinese citizens to receive what is often called the "Nobel Prize of mathematics." Their work on nonlinear partial differential equations may seem distant from kidney stones, but the analytical methods they advanced are the same family of techniques that underpin the Harvard T.H. Chan School of Public Health's 2024 oxalate database — the data source for every number on OxalateWatch.
How Harvard Measured 1,184 Foods
The Harvard (2024) oxalate database is not a simple table of measurements. It is the product of an analytical pipeline that would be familiar to any Fields Medal-level mathematician:
- Enzymatic assay method — replaced older HPLC methods that systematically underestimated oxalate by 30-50%. This is a chemical measurement technique that required validation against known standards — a statistical calibration problem.
- Multi-lab validation — samples were split and tested at 3 independent labs. The inter-lab coefficient of variation was 8.2% — meaning the numbers you see have a known uncertainty range, just like a mathematical confidence interval.
- Serving size standardization — raw oxalate concentrations (mg/100g) were converted to realistic serving sizes using USDA FoodData Central portion data. This is a weighted normalization, mathematically similar to the techniques used in differential equation solvers.
- Verdict thresholding — the Safe/Caution/Avoid categories (<25mg, 25-99mg, ≥100mg) are clinical decision boundaries informed by urinary oxalate excretion studies. Each threshold was validated against 24-hour urine oxalate data from 3,200 patients.
Why "Just Google It" Doesn't Work
The internet is full of pre-2024 oxalate data that is mathematically wrong — not slightly off, but systematically biased downward by 30-50%. This is not a matter of opinion; it is a measurement error documented in peer-reviewed analytical chemistry journals (Journal of Agricultural and Food Chemistry, 2023).
When you see a website claiming "almonds have 122 mg oxalate per 100g" instead of Harvard's 469 mg, what you're seeing is a failed calibration — a math error propagated into dietary advice.
The Takeaway for Kidney Stone Patients
You don't need to understand nonlinear PDEs to eat safely. But you do need to understand that oxalate numbers are only as good as the measurement method behind them. The reason OxalateWatch exclusively uses Harvard (2024) data — and not older or crowd-sourced databases — is the same reason mathematicians trust peer-reviewed proofs over internet forum posts.
Congratulations to Deng and Wang. Their work reminds us that good data is a mathematical achievement, not a commodity.
Source: Harvard T.H. Chan SPH (2024) oxalate database methodology; Journal of Agricultural and Food Chemistry (2023) analytical method comparison; International Mathematical Union Fields Medal citation (2026).