Hacker News new | past | comments | ask | show | jobs | submit login
I pulled data on 1378 restaurants from Google Maps to rank them in order (mattsayar.com)
77 points by MattSayar 28 days ago | hide | past | favorite | 51 comments



In most areas, this will simply show which restaurants are actively maintaining their Google Maps entry by having all negative reviews removed, which ones are putting some effort into fake reviews and which ones just don't care.

I would say the best places typically score low fours and have at least some one-star reviews written by seemingly deranged people.

One might think that caring about not having bad reviews corelates with caring about the quality of ones services, but I haven' t found this to be the case.


> by having all negative reviews removed

I've noticed the occasional incredibly bad business that gets bad reviews removed, but always wondered how they did it and how that works? A guess: they make some fake google accounts and report the review from multiple accounts?


There are companies that remove bad reviews as a service. The most popular way seems to contest that the user really was a customer, which forces Google to request a proof (e.g. a receipt). Most users either don't have the receipt anymore or don't care enough, so the review gets deleted.


There are services out there that help you do it. Like Removify: https://removify.com/

Unsure whether they're actually capable of doing it or not. But they exist.


> I would say the best places typically score low fours and have at least some one-star reviews written by seemingly deranged people.

It's heavily dependent on the area in my experience. In some places 4.3 restaurants are perfectly fine, in other it's at least 4.5. Which is why when I'm travelling I don't simply visit the first high-scoring restaurant but look closer at several ones to get a general idea of the scoring habits there.


I like to look for reviews that discuss specific menu items, good or bad. Everything else is not really helpful


Most of the time, if I want to rely on reviews for a restaurant or hotel choice, I go straight to the 1 to 3 star reviews.

If those are complains like "waiter was not nice", "they took 30 minutes to bring my food", I'll assume there's nothing to worry about.

My feeling is that bad reviews are more likely to be more trustworthy, because they are less likely to be faked (unless its the competition).


I’ve always relied heavily on Maps review but for me trust was broken few weeks ago. I went in a restaurant with thousands of reviews averaging 4.6. Food was good but not amazing, service was super kind and proactive but nothing special to deserve a 4.6 in Paris. We understood the trick at the end of the meal. Waiters kindly ask to scan a QR to make a review. You end up on a third-party landing page and if you select 4 or 5 stars you were redirected to Maps, otherwise they simply take your random-internet stars and buried them forever in this landing page. Clever.


Interesting idea for another type of score - what is the average if just taking into account 1-3 ?


I don't trust Google reviews anymore, Maps or otherwise.

Look at how many reviews will just be five-star ones, but no text. Or, one or two words quips, like, "great food!"

Negative reviews might also be emotionally driven, some by one-time events, like a to-go order taking too long because of a large office order that preceded it.

When I was apartment shopping, I saw so many "sketchy" reviews on Google and reporting them does nothing. Many would be from people who only toured the place and NEVER actually lived there, or from people who just moved in. Both are useless, but does Google care? No! Hence, my disdain for Google reviews.


It is quite time-consuming to go through the reviews due to this exact reason - 100s of low effort reviews and also fake reviews to hide negative reviews.

https://www.top-rated.online/ re-sorts the reviews (on individual place page) based on reactions and number of reviews user made. It makes it easier to see the full picture and avoid fake reviews.


We need something like Amazon Fake Reviews but for Google reviews.

Amazon Fake Reviews also stopped being reliable because people figured out how to game it.


> I don't trust Google reviews anymore, Maps or otherwise.

What do you use as alternative?


Any review site is kinda worthless.

What you can do is go through the text reviews and try to read between the lines and figure out if the place/item is good for you.


if only LLMs could be used for this... business idea anybody?


> business idea anybody?

How do you make money without skewing the results?

And can LLMs be subtle enough?

Last time I picked a restaurant via reading between the lines, it was a negative review that made me decide to go there. The person leaving it was either used to fast food like service or was in a real hurry, but between the lines the food was great, they just took a while to serve it. I was in no particular hurry so I was happy with the place.

Can a LLM figure that out? And what about the day when I am in a hurry and I am looking for fast food like service?



Friends, family, and guides like Guide Michelin for me.


Founder of Salamanca here, an app that aggregates every major restaurant booking platform into one app (OpenTable, SevenRooms, Tock, Resy, The Fork and others..)

Firstly, nice site - always love new tools to discover restaurants, thanks for posting, I’ve shared your blog post with friends, it was a brilliant read.

I have some recent experience working with restaurant reviews, I found that using only Google reviews can be unreliable, as some places that have top reviews may not be generally accepted as the ‘best’ restaurants.

We currently use a combination of Google reviews + Trip Advisor + Reviews from the booking platforms and we have web crawlers to check if the restaurant is featured on reputable restaurant guides or review sites.

We aggregate all of this review data and compute a “score”, so when users search for available tables in a city we can show available tables at the highest scoring restaurants first.

We apply Wilson score confidence intervals, to trust restaurant scores that have more reviews.

We are also applying an exponential decay when users list nearby restaurants, as you might be willing to travel a little further to go to a higher scoring restaurant.

Working with review data is fascinating.. we’re going to be launching an AI summary of recent reviews and our computed score in the coming weeks to help our users understand our ratings.

Our app went live on the App Store only a few days ago and we expect it to be live on Google play later this week.. so it’s an extremely busy time!

If you’re interested in what we’re doing please reach out, it would be great to connect, I really enjoyed your article!


It's funny seeing a Cuban place as #1, because it was the same in my old city.

The Cuban places were upscale, fancy, had amazing food and drinks. They became the de facto 'take a visitor out' spot. No qualms there with the ranking, few places even seemed to compete.

The disconnect(ie funny?) is that Americans think of Cuba as a failed, second/third world country as we're taught. So that would seemingly lead to the people/restaurants owners being cheap and scrappy. But here they are, showing everyone else up in multiple cities. I respect that a lot.


Copyright Matt Sayar©, yet:

3.2.3 Restrictions Against Misusing the Services.

(a) No Scraping. Customer will not export, extract, or otherwise scrape Google Maps Content for use outside the Services. For example, Customer will not: (i) pre-fetch, index, store, reshare, or rehost Google Maps Content outside the services; (ii) bulk download Google Maps tiles, Street View images, geocodes, directions, distance matrix results, roads information, places information, elevation values, and time zone details; (iii) copy and save business names, addresses, or user reviews; or (iv) use Google Maps Content with text-to-speech services.

Foursquare released its database of places, maybe that would be more interesting to OP (as well as the data from OSM).

Foursquare's 104M Points of Interest https://news.ycombinator.com/item?id=42219578

Foursquare Open Source Places: A new foundational dataset https://news.ycombinator.com/item?id=42191781

Demo of the dataset: https://wipfli.github.io/foursquare-os-places-pmtiles/#map=1...


It can't be illegal if it's publicly accessible content


I know that Big AI wants you to believe, but it's not true.


Hi Matt! I did something similar for https://sweetspots.fr/ it would get updated every week and was very helpful for discovering cities.

I stopped maintaining the project 1 year ago so the list are getting stale now but it was fun while it lasted. Glad to see someone else look into this!


I like how you made it into a proper shared map on Google. Happy to see someone make this into more of a product! Why did you stop maintaining it?


Well not exactly a "product" ahah more of a proof of concept. It doesn't make any money (never intended to) and I moved on to other side projects so I do not have time to maintain anymore.


All I want from Google Maps is the ability to sort by number of reviews. You can almost get it. If you click a "Busy Area", it will show you a list of businesses in that small area sorted by number of reviews and it's incredibly useful. But if there isn't a "Busy Area" to click on you can't get it, and you can't sort search results that way either.

The crappiness of the ranking and filtering options in Maps search is completely inexcusable. Ranking is Google's core business!


Someone already built a 3rd party solution for your problem:

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

Select your city and sort by "Most Reviewed." I use this site all the time to decide where to go when visiting a new city -- it’s incredibly useful.


I went through several cities I know (and several I don't) and the lists made no sense, with many misclassifications. Great intention though.


Your comment actually reminds me about a problem I had with my initial data pull. When you search a location/radius, by default Google only lists the "popular" restaurants nearby. I had to change that from the default value to "distance." Otherwise it was skipping restaurants in my search algorithm


I would rather go to a restaurant with a 4.9 rating based on 1000 reviews than to a restaurant with one with a 5 rating based on 1 review. Makes me wonder: how would that feeling translate into a ranking function though.


I've used this approach in the past with great success:

https://www.evanmiller.org/ranking-items-with-star-ratings.h...

All Evan Miller's posts on user ratings are excellent - https://www.evanmiller.org/ and search for "Mathematics of user ratings".


So what did you think of the article's approach to this issue?


Thanks for your comment, it made me do my homework. And apologies, I lazily viewed the first dozen rows and assumed that it was a simple high to low sort, and I was wrong. The author used https://en.m.wikipedia.org/wiki/Bayesian_average which answers my question. Cheers.


Author did not use a bayesian average but the Wilson score interval.



This is correct. By the way I confirmed it's the same way reddit does its "Best" sorting


Wanderlog seems to be scraping Google Maps and TripAdvisor for all their restaurant data, then rerank them.

https://wanderlog.com/list/geoCategory/1/where-to-eat-best-r...


Blog post/how he did it was discussed a few day ago (37 points, 18 comments) https://news.ycombinator.com/item?id=43026668


I want to do this where I live! I think it might be easier for me because my county has an open data portal with a list of the restaurants (because of inspections) and their locations.

Looks like Boulder County has one too https://opendata-bouldercounty.hub.arcgis.com/documents/c9d2...


Here's the source code for how I did it here https://github.com/MattSayar/restaurants_rankings


It’s super easy for where I live - there’s one restaurant within a 50km radius of us. They have no reviews. Done.


What do you think of them?


Can’t go wrong with a gigantic steak for €10, including wine, a soup starter, and dessert.


Pretty sure that's exactly how you go wrong.


lol. In many places, yes. Here, no. It’s cheap because all of the produce comes from about a 10km radius - except for the coffee, which still manages to be €0.60 for a cup.


My goodness! I’d make a special trip. I don’t suppose you’d be willing to blow the cover of their anonymity…


I find reviews in tourist places in Europe are heavily gamed by flirty waiters and begging for 5* reviews, often with freebies like a free shot etc.

I was in Spain for a couple of weeks and every 4.5 or higher rated place on Google maps was a disappointment and had handsome nice waiters

It's amazing how many people are happy with expensive poor food with nice service (I do understand service is part of eating out of course)


> I saw some shady third-party tool that scrapes the data, but it's against the Terms of Service and I don't want to worry about being banned from my entire Google ecosystem.

I'm pretty sure Google Terms prohibits using data from its Maps API otherwise than in connection with displaying/using a Google Maps service, but maybe they won't go after this, because it's not much data.


Seems the second part (well actually showing the same dataset) of https://news.ycombinator.com/item?id=43026668 Congrats for getting twice on the front page with the same material





Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: