I'm the product and engineering director at The New Yorker and this is awesome!
I'm pushing to provide public APIs, and this helps demonstrate why.
Of course, you (or anyone else here) is welcome to come and build it for real (and argue your case for/against dieresis) with direct access to the data.
If you'd like to reëngineer this for TNY, take a peek:
Hi Michael- So glad that you like it! I'd love to hear more about what your team is working on. (I suppose it's not a dieresis on/off switch.) I'll shoot you an email.
Eight restaurants reviewed in Queens in twenty years. If Queens were its own city, it'd be the fourth largest in the nation -- just ahead of Houston and not far behind Chicago.
Maybe they should rename it The Manhattanite, or maybe The Manhattanite, Preferably Below 14th Street.
Cities don't work that way though. Different areas have different mixes of residential/retail/business/nightlife.
If you slice off the more residential areas, of course you'll wind up with a big population with a smaller level of cultural activity, because cultural activity is often concentrated away from residential areas: people commute from Queens to Manhattan, and then they stay there to eat/drink. That's an oversimplification that doesn't apply to everyone... but that's the broad pattern.
For reference, according to http://a816-restaurantinspection.nyc.gov/RestaurantInspectio... Manhattan has about twice as many eateries as Queens. But that doesn't tell the whole picture, because a greater proportion of those Queens eateries are likely to be neighbourhood places, diners, delis: places that the New Yorker is obviously not going to review.
So yeah, there's probably some bias... but destination restaurants are concentrated in Lower Manhattan. That's the way it is. You should expect New Yorker reviews to reflect that.
I wonder if that's accurate. One of the big differences between Queens vs the rest of NYC is in the way postal addresses are handled.
In Manhattan every single address ends with "New York NY" and in Brooklyn every address ends in "Brooklyn NY", but in Queens the addresses correspond to the individual "cities" or rather neighborhoods that comprise the borough. So instead of "Queens NY" addresses end with "Long Island City NY" or "Flushing NY" or "Jamaica NY" despite being in the city of NYC.
So maybe this side project, which presumably is based on scraping the data from the New Yorker site, doesn't have all the various areas of Queens coded into it.
Or maybe it does, I'm just speculating wildly without actually looking into the matter.
I used the Google Maps Geocoding API to convert street addresses into lat/lng. It seems to handle the idiosyncrasies of Queens addresses well. That said, if you see something that looks out of place, please let me know and I'll correct the data.
Interesting! Red ones are closed permanently it appears. I would be interested in knowing whether there's a correlation between TNY bad reviews and whether a restaurant is now closed.
I'd be surprised if there wasn't, but only a correlation. The things that get you a bad review are the same things that stop you building up the sort of repeat custom a restaurant needs in order to stay in business.
Now, if you could show there was a causal relationship between a bad review and a restaurant closing, that'd be really interesting.
I upvoted, but it would be nice to be able to see less of a map (which forces the user to click on each dot) and maybe a list of the restaurants on the side, perhaps ranked in descending order of Yelp stars/# of reviews and with Yelp categorization that could be filtered?
A half-and-half view works fairly nicely in NY because of its oblong shape.
As another addition...you could probably pull NYT reviews fairly easily (via the article API http://developer.nytimes.com/docs/read/article_search_api_v2, or just scraping the restaurants page), and they do have a star-system of their own. I feel they'd be much more likely to issue a takedown notice, though.
edit: Though I notice you're actually using Yelp's API within their TOS [1]...i.e. freshly requesting the Yelp data with each request and not caching it...so my suggestion would be significantly harder than if you did just a straight pull of the data and organized it on your side...so ignore my TOS-infringing ideas :)
Thanks for the upvote! I really like the idea of combining reviews from different publications on one map, especially with a filter to specify which ones to make visible.
Really great idea! This is one of the first things I've ever seen on here that I think I will actually bookmark and use.
I'm just curious if you would mind divulging a little bit about how you managed to make this work? Are you just scraping the New Yorker site and sticking them onto the map, or is there something else going on?
Appears data from NewYorker is scraped (unless they have an open API?) and put into "restos" variable in JavaScript. [0] Then it's plotted on Google Map using their API. [1]
I learned: New Yorker reviewers are not very literate. Weak similes, bad puns, snarky attitude. Not really restaurant reviews; too-clever essays on their emotional ride.
No one goes to a restaurant because they need to eat. There are quicker, cheaper, easier options. A restaurant should be seen in the same light as a trip to a theatre, or to a gig. You don't go to a gig to listen to songs you could play on your home hifi, you go for the experience. The reason to go to a restaurant is because it's also an experience - the food is a part of it, but there's much more. There's the ambiance, the company, the style, the luxury, the cost.
That is what a restaurant reviewer should write about. A report on what someone ate, or what they saw, is not a review. There should be an emotional ride. That's what a review is.
If you want to read some of the best restaurant reviews ever written, search for Jay Rayner's work in the Observer. He is masterful at writing about restaurants (and food, and wine). They're a joy.
Both are true. Facts can also be useful, as well as feeling.
A useful review should include some objective analysis. Ignoring details that marred your visit but wouldn't likely impact future patrons for instance; instead of wallowing in them gleefully.
Also a good reviewer should understand and adjust for their own emotional state going into the deal. You have a bad day; you write a scathing review - how is that useful to anybody?
So, what a review is, is a lot more than disgorging a personal emotional odyssey in print. A professional would write a carefully weighed analysis including a dispassionate component to factor out their own irrelevant biases. Would include those elements that were well crafted, even if they didn't hit home with the reviewer on the particular night they visited. Would resist cheap jabs, which are the review equivalent of click-bait.
To be clear: The red pins are restaurants that have gone out of business and are closed permanently; not restaurants that aren't currently open, but will open later in the day for (e.g. for dinner).
The Yelp API provides data on whether a restaurant is still in business and is fairly accurate.
Filtering by price, date reviewed, and open/closed status are in the backlog! Would you have a preference for the price as described by The New Yorker or by Yelp? (Or by some other source?)
I'm pushing to provide public APIs, and this helps demonstrate why.
Of course, you (or anyone else here) is welcome to come and build it for real (and argue your case for/against dieresis) with direct access to the data.
If you'd like to reëngineer this for TNY, take a peek:
http://www.newyorker.com/about/careers