the whole thing, free
Open dataset
Every winner with both platforms' scores, our Bayesian-adjusted comparison, the divergence index and the closed flag from the latest live check. CC BY 4.0 — use it, cite it, build on it.
Updated · Re-ranked on the 1st of every month · next run 1 September 2026.
What is in each row
| Field | Meaning |
|---|---|
rank | TripAdvisor award rank, 1–50 as published |
name / slug | Restaurant name and our URL slug |
ward / area | Google's post-2025 ward string, and our six-area grouping |
addr | Full street address as Google returns it |
g / gc | Google rating and review count |
t / tc | TripAdvisor rating and review count |
gb / tb | Bayesian-adjusted Google and TripAdvisor scores |
score | Combined adjusted score — what our consensus ordering uses |
dc | Divergence index: adjusted gap, centred on the median of all 50 |
consensus | Agreement percentage, 100 minus the adjusted gap in points |
verdict | Our editorial stamp, derived only from the numbers |
closed | true if Google listed it temporarily closed at the snapshot |
price / price_band | Spend band in ₫, and TripAdvisor's own $ band |
family / tags | Our cuisine family, and TripAdvisor's raw cuisine tags |
Both platforms' scores go through the same Bayesian adjustment — a prior of 4.45 with weight 150 — so a 5.0 from 95 reviewers cannot outrank a 4.8 from 3,900. The divergence index is the gap between the two adjusted scores, centred on the median across all fifty. Positive means Google's larger, more local sample likes it more than TripAdvisor's traveller sample does. Negative means the reverse.
Using the dataset
Can I republish this data?
Yes, under CC BY 4.0: credit RestaurantsSaigon.com and link back. The underlying Google and TripAdvisor ratings remain theirs and are included for factual comparison; the adjusted scores, divergence index and verdicts are our analysis.
How often is the file regenerated?
On the first of every month, together with the ranking. Each file carries its snapshot month in the meta block, so you can tell two downloads apart.
What is the Bayesian prior you use?
Both platforms' scores go through the same Bayesian adjustment — a prior of 4.45 with weight 150 — so a 5.0 from 95 reviewers cannot outrank a 4.8 from 3,900.
Plan the rest of your Saigon trip
Beds, tours and transfers — the bits the list doesn't cover.
Some links here are affiliate links. They do not change your price and they pay for the monthly re-check. Full disclosure.