AI-powered website ranks San Francisco restaurants by attractiveness of patrons

SAN FRANCISCO (KRON) — Whether you use ChatGPT, Gemini or Grok, you can pretty much ask an artificial intelligence (AI) operating system anything. With the rapid rise of AI for everyday use, people have grown accustomed to asking AI anything they want an answer to.

Really, anything.

Even something as superficial and shallow as finding out where the good-looking people eat in your city — at least if you live in San Francisco. The founder of the website “LooksMapping” curated a map — using AI — that shows where San Francisco’s “most attractive diners” eat.

How “attractive” are the diners at your favorite San Francisco restaurant? You can find out its LooksMapping score by viewing the map here.

The “more red” a restaurant is, the more attractive its patrons are, according to the map. The darker blue means the “least attractive.” See the map below.

LooksMapping map of San Francisco restaurants with the ‘most attractive diners — according to AI!’

Here is where the most (and least) attractive diners are in San Francisco, according to LooksMapping’s data. All restaurants listed below either received a 10-out-of-10 score or the lowest possible score of 1.0 out of 10.

‘Top 5 Hot’

  • Ararat Kebab & Gyros (Tenderloin)
  • Himalayan Cuisine SF (Lower Nob Hill)
  • MuuKaTa6395 (Inner Richmond)
  • Hotpot Champ (Chinatown)
  • Fondue Chinoise (North Beach)

‘Top 5 Not’

  • Pizza Zone N Grill (Bayview)
  • Mandarin House SF (Bernal Heights)
  • Pier42 (Embarcadero Waterfront)
  • Asados el primo (Bayview)
  • Two Jack’s Seafood (Bayview)

The San Francisco Bay Area is considered the tech and AI hub of the world, so it’s no surprise a tool like this is being tested in The City. LooksMapping is also available in New York and Los Angeles.

New Yorkers seeking platonic and romantic relationships attend the LEGO Group’s Build-a-Bond event hosted by Tyler Cameron at The Penthouse, Wednesday, Feb. 7, 2024 in New York. (Jason DeCrow/AP Images for LEGO Group)

How did LooksMapping Get Its Results?

LooksMapping gathered nearly 3 million Google reviews from more than 9,000 restaurants across San Francisco, New York and Los Angeles. It analyzed about 1.5 million Google users who submitted a review.

Of those 1.5 million users, the study found about 587,000 had at least one detectable face. Photos of those faces were downloaded at a higher resolution then submitted to an AI tool that judged and calculated each face’s “attractiveness.” The calculation was determined by a formula (see step 4 in the document below).

Only restaurants with at least 50 detected faces were included in the map.

Full methodology and formula used to calculate a restaurant’s patrons’ attractiveness can be viewed in the document below.

The information found on LooksMapping is to be taken with a grain of salt — the website itself said. LooksMapping’s founder admits the data is certainly “biased” and “flawed.”

“This website just puts reductive numbers on the superficial calculations we make every day. A mirror held up to our collective vanity,” a statement on the website reads.

Need some dinner plans, but don’t know where to go? Maybe LooksMapping can serve as your guide to find your next tasty meal (or significant other).

Want more insights? Join Working Title - our career elevating newsletter and get the future of work delivered weekly.