Where are the best restaurants in my city? A statistical analysis

Is there an established way to take into account that people rate different types of restaurants on very different scales? "5 stars great burgers and cheap beer" vs "the avocado soufflé on the 6th course was such a letdown, no creativity at all, ruined my evening, 3 stars". I've seen particularly harsh reviews on fine dining restaurants that speak more about the entitlement of the customer than on the quality of the place.

Someone built a website that does this for every major city on earth

https://www.top-rated.online/

> But definitely don't eat at Trailer Birds.

This is the takeaway I see.

Then, as someone who's not from Colorado Springs: why is that restaurant in particular so bad? Do the reviews indicate issues with management, or issues with particular waitstaff, or issues with sanitation, or perhaps just the food and recipes are bland? In my experience from Houston, a lot of the "worst" restaurants could be seriously improved if only their management were made aware of issues that show up in reviews.

Enjoyed the read, easy to read — TIL about Bayesian average and Wilson score interval. Amazing how much LLMs helped the author achieve their goal

Best overall is just a start. Typically I have more specific needs, depending who is joining, what I had this week already. Now that you have the data, you could ask an LLM to rate each restaurant (from comments and website and menu) to find a specific recommendation:

I need a place for business lunch for 2, should be a quiet place, not pizza.

Recommend a good place for a dinner for five, should have vegetarian option. No burgers, we had that last time. We plan on sitting and chatting longer, not just food.

Nice! Would love to potentially try and reproduce this for my local town. Is there any source code that you’ve published?

Now the next step is for AI to filter out reviews written by AI.

Obligatory "highly rated doesn't equate to good" comment.