About this project
The most consensus-driven ranking of the world's best food cities

The Eatinerary is an attempt to provide the most objectively-assembled ranking of the world's most exciting cities to eat and drink in.

The idea is to help you plan your next trip, get excited about somewhere you're already going, or simply provide some escapism on a Monday morning. We want to help you answer one question: where in the world should I go next if food and drink is the priority?

There are plenty of editorial lists that rank the best places to eat in a city, and some that rank the best cities from a gastronomic perspective. The Eatinerary is not trying to be another.

Instead of curating our own list, we draw on the extensive data already available from sources like Michelin and Google Maps and combine it with more editorially focused food content from places like OAD and Eater — then explain exactly how the ranking works. Apologies in advance if you disagree with your city's score. Blame it on the data. And if you think we're missing a source we should be using, get in touch.

"Rankings are only as honest as the people making them. We try to show our working."
How it works
The Eatinerary ranks the top 100 cities in the world across five categories — Fine Dining, Casual Dining, Cheap Eats & Street Food, Bars, and Variety — using a combination of structured data from global ranking bodies, scraped editorial sources, and platform data from Google Maps and Eater.
Scores are combined into an overall city rating, with an Excellence Bonus that rewards cities scoring 90 or above in any single category. A city defined by one extraordinary dimension — Bangkok's street food, Tokyo's fine dining, London's bars — is more exciting than one that scores adequately across all five.
Rankings are updated quarterly, with the date of the latest update clearly labelled. Where real data has been applied it is marked clearly. Where scores are estimated pending full data collection, they are labelled as such. We don't pretend to more precision than the data warrants.
Guiding principles
01
Absolute counts, not per capita
A traveller visiting a city has access to its entire restaurant scene. We don't normalise scores by population. A city of 2 million with 50 Michelin-starred restaurants still has more superb restaurants to visit than one of 500,000 with 15, regardless of the per-capita ratio. To us the absolute number is more exciting — feel free to disagree.
02
Multiple sources reduce bias
No single ranking is neutral. Michelin skips whole regions. World's 50 Best has a cocktail-bar preference. La Liste overweights French restaurants. We use multiple sources in every category specifically to ensure no single list's biases dominate the score.
03
Density captures what prestige lists miss
Google Maps data — with its rating threshold and volume counts — captures bars, street food stalls and casual restaurants that award bodies simply don't cover. Tokyo's whisky bars, Bangkok's rooftop scene and Singapore's hawker centres all matter. Platform data sees them; prestige lists often don't.
04
Transparency over false precision
Some cities carry estimated scores pending full data collection. Real data, where applied, is clearly marked. We'd rather show an honest estimate than a fabricated precise number. Every score will be updated as real data is pulled.
05
Excellence deserves recognition
A city that is extraordinary in one dimension is more exciting than one that is merely good across all five. The Excellence Bonus adds up to 14 points to the overall score for any city hitting 90 or above in a category — rewarding genuine world-class performance over median consistency.
Scoring methodology
Each city is scored across five categories, each worth 15% of the overall score. A further 25% comes from an editorial weighting drawn from six global best food cities lists. An Excellence Bonus of up to 14 points rewards exceptional category-level performance. Tap any category to see the sources and weightings behind it.
Fine Dining
15% of overall
▼
Measures the absolute ceiling of a city's restaurant scene — how good are the best restaurants available to a visitor, and how many are there? Michelin star counts use a weighted points system (3-star = 3pts, 2-star = 2pts, 1-star = 1pt) rather than treating all starred restaurants as equal.
Michelin (weighted star points)
20%
World's 50 Best Restaurants + regional lists
20%
La Liste
20%
OAD (Opinionated About Dining)
20%
Eater city maps ($$$$)
20%
Cities without Michelin coverage are not penalised — the 20% Michelin weight redistributes equally across the remaining four sources.
Casual Dining
15% of overall
▼
The quality and depth of affordable seated dining — neighbourhood bistros, ramen bars, trattorias and local favourites where a visitor can eat exceptionally well without a large spend or a reservation. Covers the $$ and $$$ price range. Street food and hawker culture are measured separately under Cheap Eats.
OAD Casual + Cheap Eats (combined)
33.3%
Michelin Bib Gourmand
33.3%
Eater city maps ($$ and $$$)
33.3%
Where Michelin Bib Gourmand data is unavailable, the 33.3% weight redistributes equally across OAD and Eater.
Cheap Eats & Street Food
15% of overall
▼
The vibrancy, volume and quality of a city's most affordable food culture — street vendors, hawker centres, night markets and casual establishments at $20 or under. Google Maps volume is the primary signal; no city-size normalisation is applied. Singapore's hawker centres are classified here despite their permanent, covered nature, reflecting their pricing and cultural identity.
Google Maps (≤$20, ≥4.5★)
25%
Night & food market counts (OpenStreetMap)
20%
Time Out street food coverage
20%
OAD Cheap Eats
15%
Eater city maps ($ only)
15%
VICE Munchies coverage
5%
Bars
15% of overall
▼
Weighted heavily toward density — Google Maps at 50% captures all bar types equally, from rooftop bars and whisky rooms to dive bars and wine bars, without the cocktail-bar bias inherent in prestige lists. World's 50 Best Bars data is real (2024, full 1–100 list) for the top 10 cities; remaining cities are estimated pending data collection.
Google Maps (≥4.5★, all bar types)
50%
World's 50 Best Bars + regional lists
30%
Time Out bar coverage
10%
Drinks International Bar Awards
10%
Variety
15% of overall
▼
Measures quality breadth — how many distinct cuisine families are represented by well-reviewed restaurants. A city where 30 cuisines exist but only two are executed at a high level scores lower than one where 20 cuisines all have multiple highly-rated representatives. 23 cuisine families are recognised, with a 1.5x depth multiplier for families represented by 3 or more high-quality venues.
Google Places cuisine diversity (≥4.5★)
40%
Michelin cuisine tags (starred + Bib Gourmand)
30%
Eater city map cuisine tags
20%
World's 50 Best cuisine diversity
10%
Editorial weighting
25% of overall
▼
Six editorial sources are combined into a 25% bucket that captures the broader reputational consensus around a city's food and drink scene — the overall gestalt that individual category scores can miss. Sources span travel media, food media, industry awards and reader-voted indices to avoid any single editorial perspective dominating.
Time Out City Life Index
—
Condé Nast Traveller best food cities
—
Travel + Leisure best food cities
—
Eater City of the Year
—
World Culinary Awards
—
Lonely Planet best food cities
—
The six sources are equally weighted within the 25% editorial bucket. Where sources publish ranked lists, placements are weighted by rank. Where they publish unranked selections, city inclusion is a binary signal.
Excellence Bonus
A city that is world-class in one dimension — Bangkok's street food, Tokyo's fine dining, New York's variety, London's bars — is more exciting to visit than one that scores 70 across the board. The Excellence Bonus rewards this by adding points to the overall score for any category hitting 90 or above.
Category scoreBonus added
90–94 in any single category+5 points
95–100 in any single category+7 points
Maximum total bonus+14 points (cap)
The cap ensures the bonus can't dominate the overall score. A city must deliver across multiple dimensions — not just one — to reach the top of the rankings.
Data status
The Eatinerary is an ongoing project. Real data is applied where available and clearly flagged; estimated scores will be updated as data collection progresses.
World's 50 Best Bars 2024 (full 1–100) — top 10 cities
Live
Michelin star counts — all Michelin-covered cities
Live
Michelin Bib Gourmand counts — Michelin-covered cities
Live
Google Maps bar density — Bangkok (899 qualifying bars confirmed)
Live
Overall scores — cities 26–100
Estimated
Google Maps bar density — cities 11–100
Estimated
World's 50 Best Restaurants 2024 — full city scores
In progress
Google Maps cheap eats density — all 100 cities
In progress
Eater city maps — full scrape across all categories
In progress
Honest limitations
—
These rankings are not definitive. They reflect a methodology we believe is well-reasoned, but reasonable people could weight sources differently and arrive at different results. The full methodology document is public precisely so you can interrogate the choices.
—
Prestige lists have biases we can only partially correct for. World's 50 Best voters skew toward cocktail bars. Michelin has uneven geographic coverage. La Liste overweights France. Multiple sources mitigate this, but bias cannot be fully eliminated.
—
Estimated scores are estimates. They are built from partial data and reasoned extrapolation, and will move — sometimes significantly — when real data is applied. We flag them clearly rather than presenting them as fact.
—
No advertising influences the rankings. No city, restaurant, bar or tourism board has paid for placement or score adjustment. The rankings are produced independently.