Where in the World is Mobile Development?

In a previous post, we used the Stack Overflow Trends tool to look at how technology usage in mobile development has changed over time. But we’ve often found that popular languages, platforms and technologies differ across cities and countries as well.

In this post we’ll explore where mobile development happens, and what these countries might have in common. We’ll look at Android and iOS, and pour one out for Windows Phone.

What countries have lots of Android development?

We geocode all of our traffic and know every question’s tags, so we can plot the percentage of traffic from each country that goes to questions with the Android tag.

Android is decidedly less “hot” in the Americas, Europe, and Russia: it makes up only a small percentage (4-8%) of Stack Overflow question views from those countries. It makes up a much larger share (12-16%) of traffic in Africa, the Middle East, and Southeast Asia. What patterns could be behind this geographic distribution?

Android varies by GDP per capita

The World Bank Open Data set provides free and open access to global development data and statistics. We combined this with our traffic data to see if anything popped out. A number of indicators correlate with Android visitation, but GDP stood out as having the strongest relationship to Android development.

Countries with lower GDP per capita visit substantially more Android than countries with high GDP per capita. There are exceptions (Japan and South Korea visit more Android than the hypothesis would expect), but overall the correlation is strong (R2=0.53, p-value < 10-13).

There could be many interlocking reasons for this trend. Maybe it’s because Android phones are cheaper, or that there’s substantial Android development outsourced to lower income countries. Our traffic data can’t really answer the why.

What about iOS?

Let’s take a look at the other major mobile platform. Where is the iOS tag visited from, as a percentage of each country’s traffic?

Again, pretty cool in the Americas and Europe, and warmer in Asia, particularly China and Japan. Interestingly, we can see there’s almost no traffic to iOS questions from Africa. Unlike Android, there’s no correlation, positive or negative, between iOS traffic and per-capita GDP. This suggests that the correlation between Android and GDP may not be driven entirely by the market share of the platform in each country.

What about both?

We can combine the geographic traffic data into one plot by comparing Android and iOS traffic in each country, and an interesting picture emerges.

The first, less interesting, thing to note is that Android is more visited than iOS from literally everywhere. More interesting is that the Android/iOS traffic ratios cluster based on continent. On the lower left we have a tight cluster of American (olive) and European (blue) nations with relatively low levels of both Android and iOS traffic. Up and to the left of that we have a cluster of African (red) nations, which visit Android almost exclusively. We can also see a cluster of Asian (green) nations that develop a lot of both iOS and Android.

Conclusion

Throughout these posts we’ve been using Stack Overflow data to explore the world of software development across time and space. This is a complicated dataset with many dimensions (every country in every year has its own distinct technology profile), and it can be difficult to know where to start. Seeing these relationships between mobile development technologies and GDP gives us one piece of the puzzle. In future posts, we’ll be exploring more about how a country’s income relates to the technologies it uses, and how that can give us wider insights into the global software development industry.

In the meantime, wherever you are in the world, if you’re a mobile developer looking to take the next step in your career, here are some companies hiring mobile developers, and specifically iOS and Android developers, right now on Stack Overflow Jobs.

Author

Jason Punyon
Data Science Engineer

Related Articles

Why_Channels

Why Channels?

on
As we continue to work on Channels, we wanted to share with you our vision, the thinking behind the product and its intricacies, and insight into the process behind developing...
Continue reading

Comments

  • ondrej h

    “ios” tag is a bit problematic. Many people use “iphone” or “ios10” or “ios9” instead. Similar things happen with android. Some people mistake language with platform, therefore some mobile questions are sometimes tagged just “swift” or “java”. It’s not so easy.

    • The universe remains: not all 7.5 billion people on Earth uses Stack Overflow, so we have a reduced universe of people who use SO and tags Android/iOS. It’s just a smaller statistical universe (or population). Doesn’t interfere with the results so much, I guess.

      Definition: The population or universe represents the entire group of units which is the focus of the study. Thus, the population could consist of all the persons in the country, or those in a particular geographical location, or a special ethnic or economic group, depending on the purpose and coverage of the study.

      In that case, the universe is questions tagged Android and iOS. The result should be (almost) the same.

      • EngineerDollery

        Your definition is wrong. It’s true that most people on earth don’t use SO, but just about all mobile developers do, and that’s what this post is about. So, in this case, the ‘universe’ is ‘all mobile developers’, which means that the results are just about as accurate as its possible to be.

        • My definition is accurate. Not all mobile developers uses SO (I don’t). It’s a limited universe. It’s impossible to ask ALL people what they are doing (and, even if we could, there would be still distortions).

          The point is: it doesn’t (statistically) matter!

          • Skip

            Google the word “sample.”

    • David Robinson

      The proportion of iphone, and especially ios9 and ios10, questions is basically negligible next to questions with ios (this measures questions asked, but it’s even more true of modern traffic:

      https://uploads.disquscdn.com/images/275d9b1eabeb023dba0fe7988fa191ade300bc494d5ef381586d04b275882dbe.png

      Similarly, a search for recent questions that mention android usually have that tag when it’s appropriate: https://stackoverflow.com/search?tab=newest&q=%22android%22%20is%3aquestion

      That’s not to say it never happens (of course users make mistakes and there are subtleties in how tags are used)! However, if you adapted the analysis to take all of this into account, you’d get qualitatively the same results.

  • quantomworks [TeEm]

    Using AIR has its benefits!

  • Scott Blair

    Couldn’t it be shear coincidence with the GDP vs Android Traffic?

    • john smith

      Or higher GDP means ppl are smart enough to know iOS is a scam and Android is the same thing for cheap.

      • Yvette E

        I used to say the same thing – I use an Android phone and have jumped ship from Windows to OSX – will never go back – it was more expensive – but worth every penny. It’s a better machine. You get what you pay for.

        • Skip

          Go OSX on a hackintosh and you will be in heaven.

      • David Robinson

        Did you get the direction backwards? Higher GDP corresponded with less relative traffic to Android.

    • David Robinson

      The p-value is less than 10^-13 (one in 10 trillion), so it’s implausible that the relationship happened by chance.

      • Sure but only shows a relationship doesn’t answer the question why. Could be developing countries have more questions to ask in general related to Android mobile development.

  • Tim Castelijns

    Can you make the images click-to-zoom? It’s difficult to see read the text

  • Steven Vascellaro

    The title had me fully convinced this post was going to be about the Android Stack Exchange app.

    • 王岑

      You are wright

  • Yvette E

    It’s hardly surprising that there is a link with wealth and technologies – this is why I’ve always been kind to new comers posting in the Android tag. Android devices are more affordable and it’s free to use Android studio and learn about it. This does mean we get flooded with new comer’s questions in the tag.
    I’m acutely aware that many of us are truly privileged on the site (myself included) to be able to afford good computers, have good educations and be able to afford expensive software.
    One of the charters of W3C is to make the internet available and accessible to everyone.This is something we need to be mindful of when we see questions on Stack Overflow that we think show no effort or that the person is ignorant.
    We don’t need to keep the questions on the site, but it costs nothing for us to be kind. Behind that keyboard may be someone who doesn’t speak English natively, who lives in poverty and is struggling to get out of that, despite the odds. Be Nice- always.

    • José Bonifácio

      Can’t agree more. We have this very same phenomenon happening with [jsf] tag (surprisingly or not, it’s also Java related). Many newcomers daily, most of them not so good at english (myself included XD), I always try to be as nice as possible. Even when the question or answer is not ‘good enough’ there’s not need to be harsh. I wish everyone in this community had the same toughts as you about this situation.

  • Rocket Scientist

    That Korea and Japan stand out is hardly a surprise. Which OS is used by Korean Samsung and Japanese Sony?

    • Shahbaz Hussain

      It’s Android OS only, just that they add their own flavour and mod over it, so as to develop their own ecosystem, obviously they are companies and they need to look for their benefits.

      • Rocket Scientist

        It was a rhetorical question. It doesn’t take a rocket scientist to figure out that people all over the world are on the patriotic side and prefer phones made in their own country. Which increases demand for the OS used by those phones locally. And in turn the phone manufacturer themselves might need to employ people who know that OS. All of it creating jobs. Therefore it shouldn’t come as a huge surprise that Android is favoured in those countries.

        • Skip

          Aren’t all phones made in China? To the best of my knowledge, no phones are made in the US, Canada, Europe and Australia. I get the entire creepy South Korean/Samsung thing, but I do not think people from other countries are that strange (including Japan).

  • > Interestingly, we can see there’s almost no traffic to iOS questions from Africa
    I’m one of those “almost no traffic” guys. Do many people look for answers to their iOS questions on SO though?

    • EngineerDollery

      yes

    • Linus G.

      See how often the Objective-C, Swift, iOS, or iPhone tag are used. Quite frequently, I’d say.

  • Fuad

    There is another dimension to look at the set, like I know my country Pakistan is there both sets for simply the reason of outsourcing. We are developing android / ios most of the time for American / European market.

  • Laurent

    These maps and charts are completely unreadable because too small. About 50% of this page is just white space (on the left and right), how about using that to make the images bigger?

    • andreaiaconodisqus

      right click on the image and choose the “Open Image in new Tab” menu item to see images in full resolution.

      • Dr. Gertsen

        This is Stack Overflow… should we really have to do that?

      • Orion Edwards

        Just because a workaround exists doesn’t mean the problem still isn’t a problem 🙂

  • Ian Ringrose

    This may be related to what higher paid programming jobs (not doing mobile development) programmers can choose to take in each county…….

  • Ian Ringrose

    Making $500 from a app is nothing to me, but may change someone’s life in India.

    • mayank vagadiya

      101% truly said sir

    • e bikepk

      100% Agreed

  • inorganik

    I was expecting to see the third category, JavaScript. Clearly lots of devs are making mobile apps with ionic and react-native but this post ignores that. It’s as if you’re making the assumption that mobile development does not happen with JavaScript. Was that out of convenience, because it’s harder to separate from web dev in general?

    Further, it’s a heat map of percent of overall traffic, so cooler colors *do not* indicate fewer people visited those tags. Just that there were so many other tags visited besides iOS/Android in comparison. I think the takeaway there is that there is a broader diversity of coding languages searched in the cooler countries.

    • Bobby Drake

      You realize ionic and react-native build apps for iOS and Android right? They don’t build it for “Ionic and React-Native” devices. So why would there be a separate category for JavaScript?

      • inorganik

        Of course I realize that. The reason is because if you are developing using ionic or react-native, you wouldn’t be using the iOS/Android tags (in most cases).

  • In India, we can develop our mobile apps in cheap rate. Now a days companies want their own mobile app.

  • Yes. Mobile apps become necessary for all firms. It has made work flexible and portable.

  • Naujokelis

    My guess: developers from America and Europe have better english language skills, thus they can find an answer for their questions in few seconds. On the other hand, devs from Africa, the Middle East, and Southeast Asia have poor english understanding so everytime they encounter a problem, they instantly ask for help on StackOverflow, instead of using Google’s search. That’s just my opinion…

    • Kirk Broadhurst

      What you’re describing would be verifiable if we had data on number of StackOverflow views per distinct user or similar data. This data set doesn’t provide that information, but instead describes the percentage of traffic from a given country.

      Even if non-english speaking countries go straight to StackOverflow (maybe, maybe not) a bigger percentage of visits from those countries are mobile focused. The only valid conclusion, assuming that people from all technologies go to StackOverflow without country-tech bias, is that for every hundred developers in China there will be more mobile developers than for every hundred developers in the USA.

  • Abishek Muthian

    Are you sure the heat map is plotted right? For android development Burma glows red, for iOS it’s Cambodia. I’m not telling that there aren’t enough developers there, but from the context of the article & comparing the India, Bangladesh the heat map seems off. Can you verify the data?

  • Its time for internet. There are many websites where mobile developers are sitting online and are not getting any work.

  • I am not sure if some this data is plotted well. Esp in Africa