Hacker News Reader

Why did you include metro stations in your study? The original quote only included train stations. The metro originally largely replaced a tram network, so the metro station density in Paris is higher than for other underground systems. I think it is very hard to find kebap shops in Paris that are not within 500 m of a metro station. This kind of dilutes the significance of your study, because I never really noticed that metro stations had a negative impact on their surrounding area (the opposite, really).

I think your study would have been much more interesting (but not that algorithmically challenging...) if you only included the large train stations in Paris.

I would second this. In Paris, but I find it to be true in most European city, the area surrounding train station is usually not very nice. It is usually surrounded by cheap hotel for 1-night traveler, overpriced fast food (hence the initial theory) and social housing / cheap building (a lot of people don't want to live next to a train station due to the nuisance, so poor people often get stuck there).

On the other hand, metro station are pretty much anywhere, and they tend to have the opposite effect. Having a metro station close by is very practical and doesn't generate as much nuisance since it is underground and pretty well hidden.

I am almost sure the data will look very different if you only look for train station.

As a non-Parisian: the hard distinction between "metro" and "train" can be non-obvious, even perplexing, depending on which train systems one is accustomed with. Some systems just don't have it, in manners that if said in Paris terms it might be train to Lyon arriving next to Line 6 Westbound.

It doesn't need to be a hard distinction. I'm sure everyone can see the difference of a building with multiple platforms and usually shops and other things targetting non-locally-resident travellers and your average metro stop with has two platforms (one for each direction) and not much more than an entrance and exit. Where exactly you draw the line doesn't really matter.

That. That kind of distinction isn't universal. Some places just have a fused metro-plus-train system, in which you can slip in a $1 metro ticket, ride along changing metro trains for over a day across >1500km/1000mi, and pay outstanding fee at the exit, like on Japanese JR networks and metro trains interoperating with it.

I'm not saying which modes of thinking are more rational or superior or anything of that sort, just that the seemingly instinctive distinction you have described is not necessarily so to everybody.

There is usually a difference between a "local network" to travel within a metropolitan area and "intercity" train to go from one big city to another.

That is what matters here.

>As a non-Parisian: the hard distinction between "metro" and "train" can be non-obvious

Do you know the difference between "train" and "subway", because "the metro" is just a subway, just like in NYC or commonly found elsewhere in almost any large city.

Yeah, as a piece of knowledge, and what I've been saying is it's European thing.

I can hop onto Tokaido Main southbound at Tokyo with $1 ticket and get off at same Tokaido Main platform at Kyoto, or vice versa, or same for Ueno to snowy Aomori through Tohoku Main, if I didn't mind paying $60 outstanding fee and a 12-hour 5 train change long "metro" experience and "oh come on" look at the gate. By all definitions the Tokaido or Tohoku line are a "metro" rail, and uses some of same models of rolling stocks(cars) as literal subway services. I can technically do that because there's no hard distinctions between metro and train in the system I'm familiar with but the rails are rails and rails go everywhere.

European cities tend not to have that, and instead have often Y shaped branches off international long-distance rail networks that comes in and backs out of a "City Central" station, where it connects to local train services like trams and subways. This is due to Europe long having concept of city boundaries where city-ness of cities roll off sharp and somewhat abruptly ends, and many cities have not expanded enough to fuse together like late-game Civ maps.

Yes, I am aware, what I'm saying is that not everyone knows and you have to say it. You can't say, like, the author's stupid to not be aware of obvious distinction.

Train has a train station (with paid toilets now unfortunately in France) and connects to other parts of the country. Metro is a subway and is within city boundaries

You would need to further filter for the largest railway stations, perhaps by number of platforms ≥ 6 or 8, otherwise you will include many urban and suburban stations that serve essentially the same purpose as metro stations.

There are not a lot of stations that a Parisian dweller would consider to be a "railway station" AND "inside Paris".

I'd say mostly Montparnasse, Gare de Lyon, Gare de l'Est, Gare du Nord, and Austerlitz.

RER, Transilien, Metro and Tramway stations would not count as "train stations" in the eyes of a Parisian/French person, and places outside of the Périphérique (the ring road that follows the old defensive "Thier walls" of the city) would not count as "Paris" for this kind of conversation.

For those wondering the original image was speaking of "Gare" which are well-defined: https://www.sncf-connect.com/gares/paris

And Saint-Lazare, the third busiest train station in France based on the number of daily travellers, and the oldest in Paris.

Ah you're right! I don't know why I forgot Saint-Lazare.

I checked the link in the sibling comment[0] and just found out that "Gare Paris Bercy Bourgogne - Pays d'Auvergne" is a thing that exists. Never heard of it before.

So all in all, from a Frenchperson's perspective there are 7 "gares" (train stations) in Paris.

[0] https://www.sncf-connect.com/gares/paris

I’d go with if it isn’t Gare «something» then it doesn’t count.

This is such a dystopian phenomenon, I'm far from an expert but I guess it's a matter of city planning?

It's very hard to avoid this. It's mostly a combination of railway noise, throughput of a large number of people who are only travelling through, opportunities for pickpockets and beggars, and historically of soot and air pollution by steam engines, which makes these areas unattractive to live in.

And it is not for lack of trying. City planners since at least the 1850ies have been trying to improve the areas surrounding large train stations.

It's also a very large building or roofed space, where it's more difficult than elsewhere to accuse homeless etc of loitering.

It is difficult to avoid, as metro passengers are mainly residents or workers in nearby premises. Station halls are also much bigger and attract the marginalised who want to shelter from the elements.

I thought to include metro stations mostly because of the nature of the original French subreddit (https://www.reddit.com/r/MetroFrance/comments/1hcifh1/pas_vu...). It perhaps doesn't make much sense in hindsight, but I also filtered out metro stations later on in the post and it didn't change much.

I'm starting work on a follow-up which will sample more cities, though I still plan to use walking distances w network analysis for now.

Hey, Not sure what you were left with once you removed metro stations but in the context of the original image the "gare" would be: https://www.sncf-connect.com/gares/paris .

And people would probably often forget Paris Bercy ^^.

Seconded. Train stations are important attractors for some kinds of people and some kinds of eating place, metro stations (particularly relatively small and uniformly distributed ones, as is the case in Paris) are only able to give a restaurant an easily reachable position, only for metro travellers, compared to others in the neighbourhood. Different effects, good for a different study.

Paris is a special case - you've got a great metro and everywhere is close to a station :)

Perhaps distance to a non-metro station (eg a TER "mainline" station like Gare du Nord) would give more representative results?

I wait with interest.

That looks like a typical collider bias to me... There should be no correlation between location and quality... But as you are looking at restaurant that are still "in business" you are introducing a bias. If you simplify, a restaurant can have : - good/bad location - good/bad food

If your restaurant has bad location and bad food, it is not going to stay in business very long.

After that you can have a mix of all, but if you remove the "bad/bad" restaurant there is a correlation that appears, but it is due to the collider bias.

> if you remove the "bad/bad" restaurant there is a correlation that appears, but it is due to the collider bias.

This sounds like the correlation appears because of you throwing away some data, but the way I see it, that correlation is real - you're not removing the bad/bad restaurants, the market is.

I've been reading up on collider bias on Wiki and pondering the examples[0] - restaurants, dating, celebrities - and the way I see it, the biased statistics is still true for whoever is doing the classification (person visiting fast-food restaurants, or looking for a date), and if their selection (taste) generalizes, it might also carry over to the general population.

I feel the restaurant example from Wiki, with its associated image below, is worth discussing:

https://en.wikipedia.org/wiki/Berkson%27s_paradox#/media/Fil...

"An illustration of Berkson's Paradox. The top graph represents the actual distribution, in which a positive correlation between quality of burgers and fries is observed. However, an individual who does not eat at any location where both are bad observes only the distribution on the bottom graph, which appears to show a negative correlation."

This feels wrong to me. Why is the regression line nearly horizontal when, eyeballing the graph, a nearly vertical one would fit better and capture an even stronger positive correlation between qualities of hamburgers and fries? In fact, I'm tempted to even throw away the leftmost and rightmost points on the lower panel as outliers.

Anyway, this example assumes the bad/bad restaurants are not visited by the subject - however, if we take your scenario where bad/bad restaurants quickly go out of business, then it's the market that creates the correlation between those two hypothetically independent qualities, so as long as we're talking real world and not some imaginary spherical restaurants in frictionless vacuum, it would be fair to say the correlation exists (and that the causal mechanism behind it is market selection).

--

[0] - https://en.wikipedia.org/wiki/Berkson%27s_paradox

The correlation is still there, but it's a reason the causation you might have been thinking plausible might not be.

In OP's case, however, the correlation did NOT seem to exist.

I just want to tell you that it's very well written, well done!

> With a mighty Pearson's correlation of 0.091, the data indicates that this could be true!

I almost choked on this bit, as I was eating a sandwich (not a kebab sadly)

> I suppose I need to start working on part 2....

I'll wait for the ignobel paper.

I’d love to see more spatial sql applied. As a side effect tbh your code may get shorter. Even though R and Python are type go-to langs for ML, it is SQL which excels at spatial analysis.

Which R-Tree Rust library were you referring to?

https://github.com/georust/rstar

Banger library. Not sure how it compares to, say, kd-tree in terms of performance but it allows for plenty of primitives.

Try fitting something non-linear or simply plot a mean (or median) rating over distance with some smoothing.

please do berlin, london, and stockholm.

Yeah, this is a nice selection of European cities to compare!

Having lived in Berlin and Stockholm (and a few other cities), I'd say Stockholm has by far the worst kebab options of all places I've lived, while Berlin was the best one.

I'm curious whether this is reflected in local reviews, or whether the locals just lower or heighten their standards based on the what the average food quality is like.

Kebabs are one of the foods (like Chinese cuisine) that morphs depending on a combination of which kebab-eating group brought it to the country, and local cuisine.

I'm not sure about Stockholm, or Sweden in general, when it comes to kebabs, but kebabs in Norway bear little resemblance to the kebabs I get in London. And the UK now has a chain serving "German" kebabs, which again are significantly different from the kebabs we get from Turkish takeaway places and restaurants here... It's not that they're better or worse - I might crave one or the other depending on what I'm in the mood for, as they're almost different dishes.

So, comparing cities against each other is a lot harder than comparing differences in local opinion by location within a city.

The author is Swedish, and having asked some Swedes they basically consider the kebab to be the goto "local" cuisine.

London of course has the two styles, |Turkish| and Turkish via Germany.

Kebabs in Australia are basically a different food to the one you get in continental Europe.

Good, but different.

Berlin shouldn't be a surprise, because the Kebap Sandwich was invented here (at least according to some sources: https://berlinexperiences.com/was-the-doner-kebab-invented-i...)

Before that, Kebap was a dish served on a plate, taken in a seated manner.

The article itself gives a negative verdict on this:

  Verdict
  NO
  
  The Döner Kebab was not invented in Berlin.
  
  Popularised here? Ok.
  But there are no participation trophies for that.

And they're correct. After all, you have not experienced döner kebap until you have tasted the original Klingon jIrtaH ghab.

I don't understand how it was popularised there though. It's a staple in France and seems to have no connection to Germany

Doner kebap was invented in Ottoman empire in 19th century. It was also usually eaten inside bread.

Turkish immigrants introduced it in Germany in 70's. However it is true that the current version of Doner kebap in Germany is quite different than Turkish counterparts. (Sauces, toppings etc)

This is plainly false. Doner was being served in both sandwich and wrapped forms in Turkey well before the 70's. It was one of my grandfather's favorite foods growing up in Ankara in the 40's, and I've got family photos with street vendors selling it on the streets of Istanbul in the 60s.

It is true that in Turkey, you'll find the Döner Iskendar (cubes of bread, covered in döner meat, covered in sauce, on a plate) as a very common dish, but to claim that the street food variant was new in the 70's is insane. There are a ton of people alive today who can refute that, trivially.

As someone whose been living in Berlin for the last 11 years I'm always horrified to read people say they wish the had high quality kebabs "like in Berlin" - the average kebab here is atrocious (obviously optimized for price and not for quality) - the best ones are OK but definitely not amazing.

I don't want to find out how bad the kebabs are in Sweden and other places you hear the from like some US cities if you think Berlin kebab is high quality in comparison.

As someone who moved out of Berlin a year and a half ago — you truly do not want to find out how other people live.

I know getting something on the same level as Rüyam would be impossible; but I would pay a not-insignificant money for something resembling a random Bude near my old place.

Well, technically I did say "has best options", not "the kebab quality median is great".

I now live in Malmö, which has a much higher number of people from Middle-Eastern countries, so their cuisine is also a lot better than what you can get in Stockholm. However, since I've also become vegetarian in the intervening years I could not tell you what the local quality of the kebab is. Falafel is pretty passable though. And one of my Iraqi friends has said the better places in Malmö win the "least disappointing experience in Sweden compared to home" award, for what it's worth.

"Least disappointing experience" is how I would put the better kebab places in Berlin! Shawarma in Tel Aviv is not exactly the same but similar and so much better (of course there's a lot of variability in quality there too, and it costs something like 2-3x as much as kebabs in Berlin).

I'm sure you can also get amazing specimen in Arab countries and in Turkey (I don't have first hand experience there).

Ok but I was talking about European cities, and Döner Kebab. You're comparing it to a different dish as made in a country from the Arabian peninsula. It's fine to say that the latter is the better food but it makes no sense in the context of arguing where to get the best kebab in Europe.

I'm not arguing about where to get the best döner kebab in Europe. Just saying it's sad that the best we can get is this mediocre.

Shawarma is basically a variation of the same dish:

    "The shawarma technique—grilling a vertical stack of meat slices and cutting it off as it cooks—first appeared in the Ottoman Empire in the 19th century in the form of döner kebab,[1][14][15] which both the Greek gyros and the Levantine shawarma are derived from."

https://en.wikipedia.org/wiki/Shawarma

As someone who spends a lot of time travelling, I'd rank Istanbul to have the best döner in the world. You find alright stuff in Germany, acceptable options in the rest of Europe, and I've never failed to be disappointed by döner in the US (which is a shame, because I keep trying it because it's a top 10 good for me)

This is frankly odd, given that for most cuisines I can point at some other place surpassing the original. The best Italian food I've had has been in Chicago, the best Mexican food I've had is in San Diego. You can find extremely good authentic Chinese food in the Bay Area if you know where to look, but I can't say I've covered enough of China to confidently decide a winner there.

Yeah I can't really tell why it's so challenging to find good levantine and Middle Eastern food in Europe (even in places with lots of Turkish and Arab immigrants). Turkish döner vs Arab shawarma is more of a matter of taste (the döner is fattier and the dairy based sauces make it even fattier which is not to my taste but is not objectively bad thing) but I don't get what's so hard it making it that you can't get good results outside its homelands.

At least in Germany I suspect it has something to do with Germans considering it cheap fast food and as such are very price sensitive - there's only so much you can do when you have to cut every expense as much as possible.