Map the iPhone Users In Any City, And You Know Where the Rich Live | Hacker News Reader

Map the iPhone Users In Any City, And You Know Where the Rich Live

theatlanticcities.com

73 pointskumarski61 comments

Open articleView on HN

Super interesting.

However, iPhone appears to have been rendered last, obfuscating Android users.

Example:

iPhone only: http://d.tiles.mapbox.com/v3/enf.iphone/11/327/791.png

Android only: http://d.tiles.mapbox.com/v3/enf.android/11/327/791.png

iPhone and Android: http://d.tiles.mapbox.com/v3/enf.android,enf.iphone/11/327/7...

Sure, but why let proper data visualization get in the way of link bait?

The original source, which I'm assuming you got your images from directly, is much more interesting: http://www.mapbox.com/labs/twitter-gnip/brands/#9/40.1358/-7...

I don't see how this matches. Is owning an iPhone a symbol of being rich more than other smart phones? I know people that can easily afford an iPhone but don't have one. I know people that have an iPhone but probably should have paid their rent instead. My wife has an iPhone. We can also afford for me to have one. I choose Android instead.

It is also only geotagged tweets. I don't geotag my tweets. My wife doesn't use Twitter. So we're both completely out of this "study". It would be interesting to see stats on how many Android users geotag their tweets vs iPhone users. I believe there is a certain mental difference between iPhone users and Android users (not better or worse... just a different type of person) which could have a bias toward geotagging vs not.

Do the results take into account multiple tweets from the same person? One "rich" person with an iPhone that tweets 50 times a day is going to make a bigger mark on an area than 5 "poor" people with a Blackberry that might only tweet 1-2 times a day.

I'm not convinced there is anything useful here.

"Pew: Everyone Buys Android, Rich, Educated People Buy iPhones"

http://www.wired.com/business/elsewhere/pew-everyone-buys-an...

"Nielsen: Android users slightly less wealthy and less educated than iPhone users."

http://www.nielsen.com/us/en/newswire/2010/iphone-vs-android...

"Report: Rich People Prefer iThings"

http://techcrunch.com/2012/07/26/report-rich-people-prefer-i...

"Android Is Popular Because It's Cheap, Not Because It's Good"

http://gizmodo.com/5977625/android-is-popular-because-its-ch...

"RJI Mobile Media News Consumption Survey: iPhone users richer & better-educated than Android users"

http://www.androidauthority.com/are-iphone-users-richer-bett...

"Forrester: iPhone users are more educated and affluent"

http://appleinsider.com/articles/09/06/12/new_study_shows_ip...

He isn't disagreeing with what seems like a generally accept notion, he's disagreeing with the presentation of the data and the conclusion reached in the article.

It only consists of many geo-tagged tweets on mobile devices from many users. There are a lot of variables here without any form of control. It is a neat experiment, but you are not able to verifiably determine any kind of numbers from these maps.

Perhaps maybe there are the same number of Android phones and iPhone (and Blackberry, etc), but the iPhone users tweet 50x more - does this map take it into account? We don't know either way. What phone are they using? Maybe a lot of people in Washington are using Samsung S3s, which aren't cheap Android phones.

Bottom line, you provided nothing to the discussion other than a quick Google search without considering the many variables obfuscated within the map.

> but the iPhone users tweet 50x more

That still doesn't provide an explanation for why the data is so segregated by location.

Anyway I agree that the article's conclusion is a little overreaching, but it's still very interesting that they were able to find such a strong correlation with the data they had. Makes you wonder what other unexpected correlations can be divined from the firehose.

There are approximately 1.4 times as many iPhone tweets as Android tweets (about 1.6 million yesterday vs. about 1.1 million). It would be nice if the tweet metadata included more information about which Android model a tweet came from, but it doesn't, so all we have is the big trend.

> Perhaps maybe there are the same number of Android phones and iPhone (and Blackberry, etc), but the iPhone users tweet 50x more - does this map take it into account? We don't know either way. What phone are they using? Maybe a lot of people in Washington are using Samsung S3s, which aren't cheap Android phones.

a few outliers does not disprove the general trend presented in the article.

> I'm not convinced there is anything useful here.

Well there is indeed something here. The charts are clearly correlated with rich and poor areas (at least for cities I'm familiar with).

Now whether or not it's tied solely to iphone vs android remains to be seen. The results deserve some sort of explanation.

The charts are of dubious value. As others have send, the z-order puts the iPhone on top, occluding Android. In the "rich" areas, Android and the iPhone are close to identical.

Which, of course, makes sense, as there are high end and low end Android devices.

But outside of the "rich" areas, although iPhone tweets are rendered on top of Android tweets, Android dominates. To me this indicates that iPhones are concentrated in the "rich" areas as opposed to Androids, which are more evenly distributed. What am I missing?

You aren't missing anything. Androids are available from premium Galaxy S4s, to discount pay-as-you-go discount devices. Android is a device for everyone and every demographic, and the charts reflects that.

However by using the layering, it suddenly seems like the iPhone is the phone used in the rich areas, when in reality if you compare layers Android and the iPhone have very similar distributions in those areas as well.

One of the nice things about statistics is that cases such as those you mention tend to cancel each other out. I always laugh inside when I hear people say they give the wrong answers on surveys just to mess them up; it doesn't seem to occur to them that people with opposite points of view do the same sort of thing, and they both end up in the noise floor.

But there are things that do skew rather than cancel. So unless those are also accounted for, it is hard to trust it.

There was a thing some time ago about how a lot of phone polls are totally unreliable as a "cross section of the populous" because they really only capture the opinion of a certain set of people: Have a home phone. Are home during the day. Answer the phone from "Caller Unavailable". Does not immediately hang up on surveyor.

This is on a statistical scale.

Probably more of a hipster detector :)

How are the color pixels indicating device stacked in the map? If I only check Android I will see large green areas, but when I then check iPhone the same green areas will be red. Does that simply mean that there have been more iPhone than Android tweets in that particular spot?

I made the map tiles, but Tom did the UI and compositing, so I'm not sure how the layers are combined. There are more iPhone tweets than any other source, so they will tend to be the most visible if they are given equal priority.

And this right here invalidates the article's conclusion, the simple number of iPhone tweets overwhelms the Android tweets. The map just shows who is more chatty, versus actual income disparity.

As a side note, I do find the map very interesting - it's just the article's conclusion that I'm not convinced of.

But they don't overwhelm the Android tweets everywhere, only in some places.

I agree that it is too big a leap to say that phone choice always indicates wealth, but I do think there is a connection.

No, it supports the article's conclusion. iPhones have far more tweets, yet are concentrated in high-income areas. Androids, with far less tweets, are nevertheless more widely and evenly spread in all areas - and dominant outside of wealthy areas.

The point of the map isn't to show income disparity, it's to show the smartphone brand of tweeters. The fact that it incidentally does show income disparity is what's interesting.

You should double-check the compositing. It looks like the iPhone layer is being drawn directly on top of the Android layer. So there is a good chance that if you reverse the composting order, the map will light up green.

The correct approach would be to decide a winner, or do some blending for every pixel on the map.

Obligatory: http://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy

Being from Atlanta, this maps is drastically incorrect as to our wealth distribution. The locations on I20 East and just above 285/75on the south side have distinctly the lowest incomes and most crime. (Though, there is an amphitheater at the south hotspot)

Agreed. The Atlanta map destroys the hypothesis. The money is in the "green" norther suburbs.

I'm not seeing the same thing. The big chunk of red on the south side of the city is the ATL airport. ("Checking in! #delta #hartsfield) Georgia Tech, Downtown, Midtown, VaHi, Morningside, Druid Hills, Buckhead, Brookhaven, Marietta, Lake Claire, west side of Decatur and the south end of Sandy Springs are the areas I see in red here. These are all wealthy areas. All along the northern perimeter - which is mostly just white-collar businesses, is mostly red tinted and all along the rest of the perimeter is green. Other than Hartsfield Jackson, there is a pretty solid cone of red emanating from downtown into the north. I would say this is 100% spot on with mapping wealth to iPhones, although, I am not sold on the numbers that generated the map.

Yeah, the idea that this somehow maps to wealth is rather silly considering the data comes from geotagged tweets.

Or you can tell where university students are. Checking Atlanta, GA, you can see where Spellman college and Georgia Tech are. Check Seattle, WA, and you can see very clearly where UW is.

Also, in Savannah, GA, you can look and you'll see a lot of iPhones in the "projects".

This is a neat project, but it's not as clearcut as the author is trying to make it out to be.

In Seattle, it's interesting to see Downtown a sea of Blackberrys surrounded by iPhones. More Blackberrys and iPhones in Bellevue. Like you said, UW is another cluster of iPhones and so is Seatac airport. But the rest of the area is Android.

Tacoma looks just like the surrounding region, despite being WA state's 2nd largest city (tied with Spokane).

I don't see Blackberrys at all - those are very likely Windows Phones (considering the proximity to Redmond).

I'm not seeing it. It's a sea of red in downtown Seattle (and up into Lower Queen Anne i.e. where the Space Needle is). Yeah there's some purple in there but as it has been pointed out, surely those are Windows Phone devices.

The neighborhoods south of downtown do appear to be majority Android, which does agree with the assertions of The Atlantic.

Where was it pointed out that these purple dots are likely Windows Phones?

Maybe the shades of purple are too close for me to tell. Or maybe nobody uses Blackberrys anymore?

Cool pictures, but knowing about some of these cities, isn't this basically a population map? Obligatory xkcd: http://xkcd.com/1138/

It seems to miss very wealthy areas that have a lower population.

It's not just a population map, because you're effectively looking at the relative usage of each platform in different geographical areas. If it were a map of just iPhone users and you went: "look, all the iPhone users are in rich Manhattan" then the XKCD comic would be relevant.

It looks to me like they are drawing the iPhone pixels on top of the Android pixels... so any location that would have both ends up showing iPhone only!

Yes, but their argument is that iPhones are concentrated in the enclaves of the rich i.e. less widely distributed than Android. The way that they presented the data is actually biased against their thesis.

Doesn't work in Denver. The richest area, Cherry Hills Village, has such massive houses on relatively massive lots that there are hardly any data points, and those that are there look like an even mix of iPhone and Android.

Looks more like people tweet while out and about during the day from the patterns I see in Denver.

Looking at Albuquerque, the biggest iPhone cluster is around the UNM campus, mostly students.

I love stats. I must say, that this is a very clever metadata scraping and analysis! I do have to point out that the purpose of a mobile device though is to be able to conduct work (or the basis of this article...a tweet) from anywhere, with that said, I don't think the geo location reflects where someone lives.

Side Note: I may own an iPhone, but my wallet doesn't agree with this title.

Chicago resident here - I think the map of Chicago is fairly accurate. It's also interesting to see a number of red dots along the blue line (part of Chicago's public transit system that runs from O'Hare to downtown)

I knew it! Bushwick and Broadway near the 24-7 Popeyes/Dunkin Donuts as I suspected.

Interesting, what about the poeple who dont tweet or have a smartphone?

Also I reckon it is certainly not just income we are seeing in the ios vs android broad pattern. I would guess age and tech-savviness are a large factor in this divide aswell.

The author presupposes that rich people buy more expensive things. Some rich people are rich because they are frugal consumers.

And yet, the iPhones show up in the wealthier areas just the same.

That's a correlation.

Yes, and you seem to object to this correlation for some reason even though it's fairly clearly there.

See the comment thread above. The author didn't just present the data; she suggested that mobile phones are linked to socio-economic status.

Yes, and "linked to" is just another way of saying "correlated". I don't understand why you seem to be OK with one term but not the other.

The author didn't presuppose anything; you're projecting. The author showed you maps of the tweets of iphone users and they were concentrated in relatively wealthy areas.

You should read the whole article again. The author didn't just present the data; she suggested that mobile phones are linked to socio-economic status.

For example: "Among other things, cell phone brands say something about socio-economics – it takes a lot of money to buy a new iPhone 5 (and even more money to keep up with the latest models that come out faster than plan upgrades do)."

While it's true that iPhones cost more, they do not necessarily say something about socioeconomic status. Wealthy consumers may be frugal and purchase less expensive phones; less wealthy consumers stretch their budgets to buy iPhones.

This is making an implicit assumption that iPhones are symbols of wealth where other smartphones are not.

The real cost of smartphone ownership is in the contract. The ~$100 savings you could get in buying an Android phone (not always there - when S3 was introduced, it cost the same as the iPhone - $200) doesn't mean much when you consider $100+/month over two years. There shouldn't be a real difference in the wealth of Android owners and iPhone owners.

Not everyone has a Galaxy S{2,3,4}. "Android" phones are pretty routinely available free-with-contract and with limited plans these days. This isn't mapping high end smartphone users with iPhone-equivalent handsets. It's showing you that Apple has nothing to sell into poorer demographics, but that those demographics still want smartphones.

>with limited plans

I'm just pointing out that $2600 for an iPhone over 2 years (Verizon Share Everything, 2GB data) is only 8% more expensive than a "free" low-quality Android handset on the same plan ($2400).

Limited and prepaid plans would explain the class difference. If it weren't for the shitty suburban/rural coverage, I'd strongly consider Nexus 4 on T-Mobile.

But again: the ability to amortize a big upfront payment over a two year contract is also a characteristic of "rich" consumers. Not everyone has $200 in the bank when they need a new phone. These people end up with Gingerbread handsets.

"This is making an implicit assumption that iPhones are symbols of wealth "

It's hardly being assumed, he's comparing a map of usage with known wealthy areas and drawing that as a conclusion.

Further wealthy people can have and use things that are not symbols of wealth. They might prefer pepsi to coke moreso than the rest of the populace but that neither assumes or implies that pepsi is a symbol of wealth.

Ah, London... The City is all covered in Blackberries, as expected.

this is stupid. 1) accuracy of data 2) having an iPhone ≠ being rich.

This assumes that people mostly phone-tweet from their home, and not, let's say, from work or restaurants or clubs.

This assumes many things. Too many.

There's still a decent correlation between a restaurant's location and the wealth of its customers. People who live in Manhattan aren't typically going to go to a club in Newark or vice versa.

Anyway you can see on the map that tweets come from a huge number of locations; there's not that many restaurants and clubs.

Reply on news.ycombinator.com