Cartography Comparison: Google Maps and Apple Maps | Hacker News Reader

Cartography Comparison: Google Maps and Apple Maps

justinobeirne.com

209 pointsmanuelmagic90 comments

I am disappointed at the type of criticism of this post on Hacker News. This write-up is the kind of high-quality, original, analytical writing we so desperately need these days, in an online world that is completely saturated with clickbait that adds nothing to our understanding of the world.

Is the piece lacking in some semantics? Perhaps. But I was struck by the comments about market share and Android strategy, rather than directly discussing the article and the points it makes about the two maps at hand.

It is a rather long read about a specific topic. One could see it as a sign of respect of authority when the topic itself isn't discussed. At the very least the article isn't controversial.

I'll admit I haven't read the full article. I quit during P2. It is TMI, extremely well detailed. I'd like to address the following two quotes from P1:

> But today, with multiple zoom-levels, there’s plenty of room to show everything. Small businesses, apartment buildings, and even homes — no place is too small or too unimportant to be shown on today’s maps.

Less is more. You don't want to show as much information as you can. The user does not want as much information as you can provide. This leads to information overflow. A user wants relevant information. It is up the developers of the app to decide for the user what information they want. A way to do this, is to change it based on zoom. However, beyond that, the information could be drawn from a variety of factors, likely aided my ML/AI. For example, take method of transportation. Is the user in a car (as passenger, or driver), or is the user cycling, or is the user using the metro, or bus? Another variable could be is the user new to this environment? or yet another: who is the user? ASL? Student? Income? Coffee drinker? Interested in a coffee right now based on previous habits? Etc Google knows those details much more accurate. They're after all an advertising company specialised in data gathering company. They also may know what the user is not interested in based on profiling and behaviour. This gives Google a clear edge in this field. (But I believe Apple will follow Google's footsteps in becoming more oriented in user data -perhaps, even do a better job, too; we'll see-)

> The “POI” is dead — because today, every place is of interest.

For the developer perhaps, but certainly untrue for the user. If you take Google Maps, it appears to prioritise transit information. Apple Maps doesn't. That makes Google Maps a better use case if you're interested in public transport. Google Maps very recently (14 jan update) pushed a new version of Google Maps with native Uber support which gives them yet another use case functionality.

> Less is more. You don't want to show as much information as you can...

100% agree, but I don't think it's saying more is better.

Rather that because maps now have dozens of zoom-levels (as opposed to just 1 with paper maps), they can show all sorts of things that wouldn't have been included before. Things like houses, small businesses, etc. Things that would've never been on a paper map.

This makes the POI dead because now anything can be included, anything can be on the map (regardless of its importance).

(Example from my own experience: Google maps has an outline of my house on it and even my backyard tool shed. But I've never seen a paper map that also had this.)

Also 100% agree with you that personalization is awesome, but it seems like the whole point was specifically to examine the defaults. To see what decisions were made for the defaults and how it made the two different.

> Rather that because maps now have dozens of zoom-levels (as opposed to just 1 with paper maps), they can show all sorts of things that wouldn't have been included before. Things like houses, small businesses, etc. Things that would've never been on a paper map.

Yes, I got that point, but I think my counterpoint was stronger.

Perhaps I'm odd, but a POI which doesn't have my interest is not a POI. If I'm in Paris, and I have no interest in seeing the Eiffel Tower, then the Eiffel Tower is not a POI for me. It may very well be one for other people, and I respect that, but I've seen the damn thing on multiple pictures already (I've seen it IRL as well).

So from a technical PoV the POI is dead because everything can be a POI? This suggests the developer believes more is better. Whereas when I looked at Google Maps reducing the amount of information I get the impression they realise less is more.

The default should not matter. Do the default search results on Google Search matter? Given the rise of the bubble, less and less. I hope the same can be true for Google Maps and Apple Maps. At the same time, we must be careful to regularly think outside of our bubbles, or look into the bubbles of others (e.g. your wife's, son's, aunt's, etc).

> I think my counterpoint was stronger.

No, what I was saying was that your counterpoint isn't a counterpoint. A counterpoint would be: Google shouldn't show new kinds of things on the map (such as houses, small businesses, etc.).

The OP is talking about new kinds of things, while you're talking about the density of things. And the OP actually makes the same argument you're making a few paragraphs later:

> we’re also seeing the argument for map personalization: Not everyone uses transit, so there might be places—such as landmarks—that are more important to some people. And when it’s life-and-death, hospitals are the most important places in the world. But how often does the average person visit a hospital? Personalized maps are better at surfacing the most appropriate places for each person; but for the default map, something still has to be chosen… and that’s what we saw above.

Personalization is great, but something still has to be chosen for the default. That's why the default matters.

I just wanted to say thank you to author for this incredibly high quality, well done piece.

Looking forward to reading the next parts!

It seems that google has a more coherent theory of how users use maps at different zoom levels.

At the outermost zoom, physical geography is emphasized : terrain shows prominently, and political boundaries at the national level with hints of state boundaries.

Zoom in farther, and political bondaries are more emphasized and physical features less. States are named and some cities.

Zoom in farther and you start to reach the domain of navigation. Large roads are prominent and labelled, as are some cities.

Zoom in farther and more roads are labelled, as well as rail/metro stops and some landmark destinations.

At the highest levels of zoom, finally it transitions to destinations and foot-navigation as the focus.

Apple has a similar progression, but with less coherence. It emphasizes cities/neighborhoods more at the expense of road labels. And landmarks and destinations appear at lower zoom levels, also at the expense of road labels. Rail/metro stations are much less prominent.

Alternative: OpenStreetMap !

"The following companies offer consulting, tile-hosting or other services for sites wishing to switch to OpenStreetMap."

https://switch2osm.org/providers/

* http://www.weeklyosm.eu/

* http://www.openstreetmap.org/help

* Tips for new (Pokemon Go) mappers: https://blog.openstreetmap.org/2016/12/30/tips-pokemon-go/

From a cartographic perspective I would love to see this expanded to include Mapbox, and Bing. As well as Yandex, and any other foreign players whose names I cannot remember right now. I would probably be more interested in seeing how they compare, instead of just US based ones.

Here WeGo (formerly Nokia, originally NavteQ, now I think owned by BMW?) would also be interesting.

EDIT: Turns out it’s VAG, BMW, Daimler, Intel and some more.

Here is owned by a consortium of automakers and recently, Intel and a couple others.

https://en.wikipedia.org/wiki/Here_(company)#Acquisition_by_...

I am still working my way through the whole thing, but so far it seems like the differences between the two are in focus. Apple seems to focus on the Where aspect of the map with more city labels and such. Whereas Google is focusing on the How, meaning how to get from place A to B. It focuses on the roads and means of transportation.

I wonder if there is a different focus in Russia or China, or anywhere else for that matter. Are there more ways to focus or do map providers always have to choose between these two options?

I certainly prefer the Google method, but I do not know if that is because I am used to it, or if it really provides me with more value. Does anyone prefer the Apple maps approach? And if you do, is there any specific reason? I am truly fascinated by these differences.

From a cartographic perspective judging Mapbox is difficult as it is a platform that allows vast customization of your map on a user by user basis.

Its true, I was thinking about that same thing. But they still are making some high level decisions. Would definitely have to compare the default, and avoid the custom ones.

> And both are in a race to become the world’s first Universal Map — that is, the first map used by a majority of the global population

This is a very US centric view.

The comparison is done in the city where the two companies are nearly located, the other city is the most important in US and the other is the most important of the English speaking cities in the world outside USA.

If you want to talk about global mapping, use global comparisons.

And not talking about the fact that Apple Maps are supported in just one platform and the market share in USA is very different than globally.

This. Both map apps are very good in the states but suck in other countries.

For example, Google maps can't even show directions in South Korea while Apple maps can due to government restrictions.

> And not talking about the fact that Apple Maps are supported in just one platform and the market share in USA is very different than globally.

Agreed, Apple disqualifies themselves from the race to make the Universal Map.

It is not a race to become the universal map. When Steve Jobs was breathing, there were a few fierce words about 'Android'. Having Google Maps or a Google browser on the phone was to be avoided at all costs so Apple shopped for maps and put something together so they would not be at the mercy of Google deciding to discontinue their maps app on iOS/Safari. So the Apple maps were never born out of vision, just spiteful thinking and a perceived need to have indigenous maps that were not supplied by the old sparring partner (Microsoft) or the evil Google.

I think that the Apple map product was designed in that era of skuomorphic thinking that went on until recently in iOS. So the maps are designed to be a familiar map, much like the atlas you had as a kid when it comes to typography.

The Google Maps evolved from looking like paper maps a long time ago and are about showing you the information you want, not everything that can be crammed on screen. This challenges us to think about how we see the world, for a long time 'A-Z' maps were the thing and you had page numbers and 'square' numbers, e.g. B7, page 126 instead of 'hyperlinks'. Apple maps are more in line with that 'old persons' way of dealing with cartography issues, the Google maps are pioneering a better UX for visualising the world.

It wasn't spite. Google introduced a large new feature as Android-only (turn-by-turn directions) and refused to put it on iOS. So this wasn't just a fear that Google would use its position to screw over Apple, Google was actually actively doing that.

No, they didn't refused, Apple and Google didn't agree in the terms to put the feature in iOs.

The maps apps was developed by Apple

Where are you getting your information? Because that's wrong. Google introduced turn-by-turn on Android and deliberately kept it out of iOS because they were trying to make Android more attractive. Google and Apple had an existing contract at that time, there were no terms to negotiate. And so when the contract did expire, Apple dropped Google Maps and introduced their own Apple Maps because Google had already demonstrated their willingness to screw over Apple.

The app was done by Apple, not by Google. Apple used the Google Maps API and that API doesn't allow turn by turn.

So no, Google did not screwed Apple, both companies didn't agree to for the terms to use the turn by turn.

That's all, none of them were guilty of anything.

The app's code was written by Apple, yes, but it was done so in close collaboration with Google. It's not like Apple was just another API consumer. And if Google hadn't had their heads up their asses, they would have given turn-by-turn to Apple and perhaps enjoyed still being the default mapping app on iOS.

Also FWIW it wasn't just turn-by-turn, it was also the fact that Google wanted Apple to give them more user data, which Apple didn't want to do. So even with turn-by-turn it's possible Apple may have chosen to do their own maps anyway, but turn-by-turn was definitely a major reason why Apple dropped Google and built their own service.

My God, I have said that both didn't agree in the terms, Apple wanted turn by turn and Google wanted more user data for giving it.

But yes. Evil Google sdcreweed Poor Little Apple.

They had to give ALL the features without receiving anything, who they tought they were?

You are aware that Google makes a large portion of its mobile profits from iOS, right? Even though Google wasn't getting as much user data as they wanted from Apple, being the default mapping app on iOS was still a big benefit for them. It's not like Google was being generous by providing maps for iOS. So Google just screwed themselves. Google thought they could force Apple to do what Google wanted, because they thought Apple didn't have a reasonable alternative. But it turns out that Apple did, which was to build their own independent mapping service.

> You are aware that Google makes a large portion of its mobile profits from iOS

And you can provide a source for that claim, isn't?

But yes, we are aware that you're convinced that Google were just a evil and stupid company that screwed themselves.

In what world two companies can disagree, Apple was totally right demanding everything without giving more.

Problem is that reality is what it is, both of them are powerful companies, both of them have their interests and they didn't aligned. Nothin less, nothing more. Not he fracking black and whit world you paint

I really don't know what kind of argument you think you're having, but each new comment is making less and less sense. I also don't appreciate how you keep trying to put words in my mouth.

The whole point of my original comment was to just dispute the claim that Apple developed their own Maps out of spite. I don't know why you're trying so hard to pretend that the removal of Google Maps was entirely Apple's fault, and why you keep repeating the same claims that I've already debunked (e.g. "Apple was totally right demanding everything without giving more").

The article mentions that the U.S. focus was deliberate in order better assess each map's priorities:

"we want to look at areas where the maps are most likely to be “finished” (i.e., areas where the maps are most likely to be complete and up-to-date.) San Francisco is Google’s and Apple’s backyard, so the maps should be really good there. And New York and London are two of the world’s largest, most important cities—so they should’ve also gotten a lot of attention."

> If you want to talk about global mapping, use global comparisons.

Worth pointing out that the title was "Cartography Comparison: Google Maps and Apple Maps", not "global mapping comparison".

> And both are in a race to become the world’s first Universal Map — that is, the first map used by a majority of the global population. In many ways, this makes Google Maps and Apple Maps two of the most important maps ever made.

This is the second paragraph

I strongly prefer apple maps.

Why?

Because I have low vision and only Apple Maps respects the system font size. I can actually read the map labels. Google maps renders the print too small.

If I can't read the map labels, I can't use the map. Everything else is gravy.

I wish there were more settings and personalisations available in both Maps apps. I’d like to activate an option that always shows me all public transport stations, regardless of zoom level.

There's a reason why it is not being done - you wouldn't be satisfied with the speed. Try some old-fashioned GIS application, that queries and displays exactly what you have selected and compare that with Google Maps.

These map apps with cached tiles (whether raster or vector, doesn't matter) gain speed precisely because they limit what the user can choose and then cache that. The normal GIS apps cannot do that, but conversely, their users are complaining that Google Maps is much faster...

I would think that vector tiles + display filtering would be roughly just as fast as vector tiles without any filtering.

At least assuming a halfway decent implementation (it shouldn't take real long to process the few hundred items that would be present in the viewport).

You certainly could filter what you receive in the vector tile. However that vector tile is already a cached / filtered piece of the data. It limits what you would be able to do.

Sure, the client can only filter the data it receives. But the tiling approach and client side filtering can clearly be applied to POIs, and while I haven't actually experimented with such a thing, I would expect the performance to be quite good even if the user had lots of filters set.

Correct me if I'm wrong, but doesn't MapBox with vector tiles essentially cache and serve the raw GIS data (in PBF format) on a tile by tile basis, and then render that on the client?

No, it's not raw GIS data. Even the vector tiles are generated for a certain scales (or set of scales). They contain only data that should be displayed at given scales and also the geometries are being generalized depending on the zoom level.

What Mapbox can do, is to _remove_ data that are not in stylesheet - but as I understand it, that is being done server side (https://www.mapbox.com/blog/style-optimized-vector-tiles/). So in theory you could put more data into generated tiles, users could create their customized stylesheet (would have to be found out how, the server needs to have it accessible, it does processing based on it) and it could work - for a few of them. If everyone would do it, you are back to square one. This is usually usable for a defined roles (what group can see what in your data), not for picking and choosing every possible object at arbitrary scale.

The thing is, you usually cannot put all the data in the tiles and just serve them - most users won't use them anyway, so you would just waste bandwidth. That one guy who wants something special usually wants other object than that second guy, who wants also something special.

Mapbox does maintain a set of (more?) complete OSM vector tiles, aimed at quality assurance tasks:

https://www.mapbox.com/blog/osm-qa-tiles/

How much bandwidth is actually wasted? i.e. if you were to put all data that's currently in OSM, just chunked into tiles, how much bandwidth would be required to serve it?

It would seem to me that, as pipes get thicker and bandwidth keeps growing, we will inevitably hit the point where it's easier to just server raw data and let the client render it soon enough.

Mapbox allows you to customize tile sheets as per your fancy. You can even select what symbols will appear at a given zoom level.

[1] https://www.mapbox.com/help/customizing-map-colors/

You can do this in OSMAnd.

It has a rather complex UI though and the availability of POIs will depend on the quality of OpenStreetMap in the area you are interested in.

I wish they worked offline. I was in Tokyo in 2016 and Google Maps was useless without internet.

Google Maps works offline. You can select a zone in a map and download it. Although it's not perfect and has quirks like not being able to plan a walking route offline.

Not in Japan, because of IP rights issues. Not all regions are cacheable.

It doesn't work in Myanmar either I have found. I'm guessing more for politic than IP right issues though.

Does the transit layer in google maps (hamburger menu->Public transit) work for you? It doesn't _always_ show all transit stations regardless of zoom level because, to be blunt, that's impossible (degenerate case: zoom out to the entire world) but it definitely prioritizes them at lower zoom levels.

The article reads like there are no noteworthy alternatives to those two. But in my experience, at least HERE Maps and Bing Maps are quite up to par. Especially HERE is quite strong when it comes to navigation.

They lead with this, which is probably why others aren't considered...

>> both are in a race to become the world’s first Universal Map — that is, the first map used by a majority of the global population

Is Apple actually trying to have a universal map? Can you even use Apple maps if you're not on an Apple device?

Exactly. There are a number of maps racing for world domination, Apple isn't one of them.

Except that HERE has more of the world mapped than google...

>> But in my experience, at least HERE Maps and Bing Maps are quite up to par

when I had my Windows phones, I preferred the Bing Map app to Google Maps. I always thought the UI was superior and more user friendly to Google Maps.

I had a similar experience on WP and preferred HERE Maps.

One reason for the poor quality of Google Maps was the lack of an official App, and third-party ones were quite poor. Google seems to have a policy against releasing apps for Mobile Windows platforms that goes further than just "we don't care".

Since Windows phone has such a low % of users and then expense in supporting some thing as complicated as Google maps it would make sense to focus on the two. With Android and iOS they can hit 90%+ of users.

As an engineer, I don't want to build something for a 0.3% (and declining) platform while my colleague is building the same app for a 80% platform in the same team. It's like you are writing codes which will be soon deprecated.

EDIT: market share http://www.idc.com/promo/smartphone-market-share/os;jsession...

Yep, and now you'll probably find them in cars as branded built-in nav.

Wow, that would be nice. I remember having a MUCH better experience as well when my Windows phone paired with the Microsoft Connect in my Toyota Corolla.

The difference between the Windows phone and Android was pretty huge:

Android text message:

- Message pops up that I have a text message, I have two buttons, ignore or read.

Windows Phone text message:

Voice alert asks if I want to read it or ignore it. After reading it I can:

- delete it

- respond via voice command

- or call person back from number

It's a pretty big difference, and one I really noticed when I went back to my Android phone.

Yeah, sadly that doesn't mean it has the fully same UI. But at least in my Mazda (which uses HERE as the nav driver), the UI is noticably better designed for driving than the phone interfaces are. Minor details mostly, but noticable.

> Especially HERE is quite strong when it comes to navigation.

IIRC, HERE maps is built by actual navigation rather than crowd sourcing which makes it more accurate than other ones.

Google did use their Street View cars to check navigation. They also use government data and crowd sourcing. I imagine that, on average, they have the best maps in areas where they are active. HERE may have better coverage overall.

In areas with active communities, that OpenStreetMap is crowd sourced is probably more of a strength than a weakness. The issue for OSM is that there are lots of places where there aren't any active contributors, never mind a community.

This tools makes it easy to compare several web maps:

http://tools.geofabrik.de/mc/

My only real complaint about HERE is that it's search is inferior. Google maps always lets me just type in the name of the store or restaurant, while with HERE I find that far too often I have to use Google to find the actual address and then type that in.

In my Windows Phone I really LOVED HERE Maps, but I'm quite disappointed about Maps on Windows 10 (I own a Lumia 830 and I updated from Windows Phone 8.1 to Windows 10 Mobile).

I know that Maps use data from HERE, but the search algorithm it's not as good as on HERE, and until a few months ago the Maps user interface was rubbish. Now it's a better but still there is no alternative to Nokia Drive on Windows 10 Mobile: you simply can't (or at least I didn't find a way) to activate the drive interface without looking for a destination.

Windows Maps is less clever than HERE, for example if I look for a street named "Via Giuseppe Garibaldi, 4" (I'm Italian) often it doesn't show the nearest one to me, but it drives me some other road, often in another city. And I always need to remember to input the exact name of the street, if I just digit "Via Garibaldi, 4" (very common form of abbreviation in Italy) it doesn't show any results or it gets me a completely useless one. HERE worked all the times. It is also not uncommon that number positioning on streets is not accurate; sadly, there were times I had to open Google Maps on Edge to look for the correct position on #34 on the street I was on, because Windows Maps was showing incorrect data and I couldn't find it by myself (that is the reason I was using a GPS navigator in the first place).

There are alternatives. But what is the marketshare of the alternatives? And how do they compare?

Does HERE have anywhere near the marketshare of Google Maps (>1 bil.) or Apple Maps (~800 mil.)?

Even though they have their own app, HERE is primarily a data provider for others (car manufacturers, I think Bing maps)

Ok, a very interesting way to analyze maps. Like a lot of engineers I enjoy different maps. It wasn't until we replaced our wall sized US Map this year (we got a large format US Map, World Map, and Globe when homeschooling our kids) that I realized how different maps could be (both good and bad).

In my fantasy science fiction imagination I've got a map on an large format HiDPI screen that I can drop a Surface Dial on to and turn it to pick out the features I want from Pilot to Topographic to Road Atlas to Political Districts. We have the bandwidth and storage to do that now, just need a designer and business case.

Small discussion from when the article was newer: https://news.ycombinator.com/item?id=11836584

While specifically cartography, the recent trend to 3D render satellite views by both Google and Apple is pretty bad, and a step backwards. Both companies try to project the 2D photograph onto a 3D mesh, which fails to work in many cases due to alignment and measurement errors. Instead of getting nicely wrapped polygons, you get weird crumpled shapes with distorted pixels projected on them. While annoying, this wouldn't be a show a stopper because you can always click "2D" to see the undistorted 2D photograph right?

WRONG! 2D simply rotates the 3D model so that the z-axis is perpendicular to your screen. All the distortions and artifacts remain in place! This is broken.

This is so awesome. High quality, in-depth, and very on-point. I love whoever went to all the trouble to make this. Awesome content!

All of the example images are too low resolution to actually make up your mind about what the author is talking about when referring to them.

The images enlarge when tapped upon

Not by any meaningful size on my machine.

The image loading on this site was pretty unbearable; the images loaded from the bottom up lazily and were pretty low resolution.

I ended up loading all the chapters before I read the 2nd page, so the rest of the pages would be ready for me. I don't know if the image loading was a design choice or just because of the HN traffic.

Very fascinating topic.

Agreed, came here just to make this comment. It's silly that a piece comparing the two map apps - inherently very visual - has such bad imagery, especially considering the author clearly spent good time making those images. Worse, because of the image loading design I can't easily zoom. Would have been better there was no image loading and I could simply use Chrome!

All of the images enlarge when you click on them. (It's mentioned in the Intro.)

[May 2016]

Thanks O'Beirne! That was really interesting.

Off topic but I found this interesting:

Out of the top 10 mobile apps in the US 5 are owned by Google, 3 by Facebook, and 2 by Apple.

> Who will get there first?

Isn't it pretty clear that Android is well ahead in global market share? I mean, Apple isn't even trying to become the market leader, they are solely focused on the premium segment.

I stopped reading right there.

If you live in a bubble that small I am probably not going to learn anything from reading your article...

i am more interested in accuracy comparison than i am in design.

That's in there, see part 2.

> And both are in a race to become the world’s first Universal Map — that is, the first map used by a majority of the global population

Actually, lol, no. OpenStreetMap is. Google Maps will be a contender for some time, but eventually it'll settle.

Then why does nearly every business use Google Maps and it's API for map features? You have to pay for maps access on a commercial basis, so they'd be stupid to ignore a free (and according to you superior) solution.

My guess is you're a FOSS zealot who can't admit that a commercial rendition of a particular idea is superior, because it would crush your ideologies and beliefs.. (and I say that as someone who has no problem with using good FOSS like VLC or Transmission)

Many businesses use Mapbox (which is based on OSM, but is not free) and many consider that superior choice to GMaps. Others use Bing maps (Facebook, for example). Then there are Here maps, Yandex maps, etc.

GMaps are a quick and comfortable choice, but if your business really depends on communicating with maps, or you want to do something more than display pins or overlay kmls, it is not a best choice.

My boss was complaining that COTS were expensive and gave gnat as an example. I asked him why do we pay for it ? We use support less than once every 5 years and in case of real problems, public forums exist.

He was completely against this idea. For the company, paying for support is like having an insurance, they do not want to drive without insurance.

I suggest trying OSM, at least. It really is a superior solution to both Google Maps and Apple Maps in many respects, especially when it comes to level of detail and how up-to-date it is.

Google Maps is clearly in that race, and I assume leading. Apple Maps clearly isn't unless they plan to at some point either build a product accessible to the majority of the global population or make a version for a platform that cares about that, like the web or Android.

2016 Nielsen for popular apps had Google Maps #4 for all apps. There were no other mapping apps on the list this year. Not even Apple maps so talking about Google maps as the universal solution makes sense, IMO.

This article refers to 2015 Nielsen when Apple maps did make the list. Not sure why it fell of the list.

Reply on news.ycombinator.com