05-07-2025
AI map blasted for ranking restaurants with attractive customers
A new interactive map claiming to rank the 'hottest' customers at thousands of restaurants in New York City is facing backlash. The project — created by 22-year-old San Francisco-based programmer Riley Walz — uses AI to analyze millions of Google Maps reviews and assign attractiveness scores to diners. His website, LooksMapping, aims to reflect what he calls society's own superficial habits. In his latest project, titled 'Finding Which Restaurants the Hottest People Go To,' Walz processed around 2.8 million reviews from 9,834 restaurants across NYC, Los Angeles, and San Francisco. The data powers a heat map that highlights venues based on the perceived attractiveness of their patrons — sparking controversy over its premise and ethics.
Now, the 'official' results for the Big Apple are in. Urbani Midtown takes the top spot, ranked as the restaurant with the most consistently 'hot' diners. At the other end of the scale, Jimbo's Hamburger Palace in Harlem came in last, labeled as attracting the 'least hot' foodies. While the rankings may seem shallow, the site isn't just a beauty scorecard. Instead, it aims to reflect how society — especially in major cities — assigns value based on appearances.
'The model is certainly biased. It's certainly flawed,' Walz himself wrote on the website. 'But we judge places by the people who go there. We always have. And are we not flawed?' he added. 'This website just puts reductive numbers on the superficial calculations we make every day. A mirror held up to our collective vanity.' The one-dimensional interactive map spans the entirety of New York City's grid, displaying thousands of color-coded pins strategically placed across the city's vast network of restaurants.
It uses a scale from 1 to 10 - with 1 representing eateries frequented by the 'least hot' diners, marked in shades of blue, and 10 representing those with the 'hottest' patrons, highlighted in firetruck red. By clicking on any of the pins, users can view a restaurant's rating along with three additional metrics: a scale indicating how 'hot' its diners were ranked, an age range distribution of patrons and a gender breakdown showing whether the crowd skews more male or female. But it raises an important question: how exactly was it determined who qualifies as 'hot' or 'not'?
Walz used a computer model to scrape 2.8 million Google Maps reviews, isolating reviewers whose profile photos featured a detectable face - 587,000 profile images from 1.5 million unique accounts in total, as reported by The New York Times. Those profile photos were then analyzed using a set of descriptive phrases designed to assess attractiveness, as well as age and gender. The phrases included: She is attractive and beautiful, he is attractive and handsome, she is unattractive and ugly, he is unattractive and ugly, a young person and an old person.
Relative attractiveness scores were then calculated for each reviewer, though Walz confessed to the NYT that 'the way it scored attractiveness was admittedly a bit janky'. The AI model appeared to favor superficial or arbitrary details when gauging attractiveness - for example, a profile photo of someone in a wedding dress might be rated as 'hot', while a slightly blurry image could result in a lower score.
'The model isn't just looking at the face,' Walz told the NYT. 'It's picking up on other visual cues, too.' According to the data, the top five Manhattan restaurants with the 'hottest' diners -each earning a perfect 10/10 - are Urbani Midtown in Midtown East, Shinn WEST in Hell's Kitchen, KYU NYC in NoHo, Aroy Dee Thai Kitchen in the Financial District and Thai 55 Carmine in the West Village. Urbani Midtown describes itself as offering 'the best Georgian food in a casual and friendly atmosphere that will make you feel like you are eating in Georgia', according to its website.'
