Comment by jack-r-abbit | Hacker News Reader

Comment by jack-r-abbit | Hacker News Reader

Parent Full thread

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.

> 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 :)

"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.

Reply on news.ycombinator.com