Highlights from the Stack Overflow 2016 Developer Survey

For the 56,033 developers who took the 2016 Stack Overflow Developer Survey, and the rest of us who wanted to see the interesting stats without having to do any work, the wait is finally over! The fancy infographics for Stack Overflow's 2016 stats are here and it's time to dig in! If you just want the cliff notes for now, read on and I'll give you some of the greatest hits from the survey.

Star Trek vs. Star Wars

I have to start with the most important data this survey collected—whether developers prefer Star Trek or Star Wars. While 'Wars blew 'Trek out of the water among developers under 50, Trekkies won in the 50-59 and 60+ age ranges. Maybe because those older groups saw the original Star Trek first, before Star Wars even existed. I think it's important to note that Firefly was the top write in, and it is also the top show in my heart.

Now for the real survey question: Kirk or Picard? (or Sisko or Janeway)

91% of developers are employed, but 78% are open to new opportunities

Don't count on loyalty to keep your developers from leaving. Bean bag chairs and foosball tables don't make up for non-competitive salaries. 63% of developers included salary on their list of priorities, which was the highest percentage (I'm rounding to the nearest percent for most of these stats throughout this article).  Work life balance was second behind salary with 50%.

Sure, after a certain salary level, the money becomes less of an issue. 37% didn't even include salary in their priorities. It also depends on the country you live in and the occupation too. But money is always tempting.  Don't take your developers for granted.

Stack Overflow Developer Survey 2016 Results https://t.co/QQS95UKzNn // Money buys happiness buys money pic.twitter.com/OZrq2z1lgk

— enguerran (@ticabri) March 17, 2016

Another interesting stat: referral by friends was the most common way respondents discovered jobs, except in India, where in-house recruiting was the main method.

Self-taught developers went way up

Last year, 42% of developers said they were self-taught. This year it's 69%. How does that happen? Maybe Stack Overflow changed the question this year and didn't say anything. This was a select-multiple question this year, so it doesn't just mean that the developers only learned their coding skills through self-teaching. That makes me wonder why the number isn't higher.  Surely every developer learns to use some tiny library by themselves at some point in their career.

Another interesting stat: only 45% of the surveyed developers have a college degree in computer science or a related field. This makes sense since Google interviewers don't care whether you have a degree or not.

Three years into a Programming degree and I'm starting to think I may as well just give it to Stack Overflow when I finish.

— Gav (@GavMakesGames) March 10, 2016

"Trying to be nice" becomes less of a challenge as you gain experience

When asked about the biggest challenges at work, respondents said "unrealistic expectations" the most often (35%) while "poor documentation" (35%) and "unspecific requirements" (33.5%) were close behind. Developers feel these challenges increase as they become more experienced.

One challenge that goes down slightly with experience is "trying to be nice".  It's not a challenge if you stop caring about it right?

Please consider attending my unsession "Let's All Point and Laugh at this Ruby Dev Trying to Code Clojure for 30 Minutes" #fakestrangeloop

— Neckbeard Hacker (@NeckbeardHacker) September 19, 2014

Everyone is "full-stack" now

While it's an exaggeration to say everyone is full-stack, it's the most common developer occupation by far among survey respondents.

By far the most common developer occupation is "Full-Stack Web Developer" with twice the responses of next category https://t.co/8r9fGcUMDX

— Peter Stone-Thompson (@pwdst) March 17, 2016

28% called themselves full-stack developers, and the runners up were back-end developer (12%) and student (11%). Obviously this question had a lot of diversity in titles. However, I think it's fair to say that full-stack isn't just a buzzword used to make yourself look good in an interview. According to the SO analysts, full-stack developers were comfortable using 5 to 6 major languages or frameworks, vs. 4 for all other occupations.

JavaScript all the things?

JavaScript is now the most popular BACK-END technology, even beating SQL (Java, PHP and C# follow). https://t.co/8r9fGcUMDX

— Peter Stone-Thompson (@pwdst) March 17, 2016

This is very surprising. It's actually kind of hard to believe that this stat might reflect the entire software development community. 

I think this is where we might be seeing the skew of developer stats that come from Stack Overflow. At SO, you ask questions. Which means the active community may actually be composed of either less experienced programmers or developers that are working in a language ecosystem that is new to them. JavaScript and Node.js apps are the fastest emerging ecosystem in software right now, so it makes sense that JavaScript has the most tags on SO and that most of the respondents in this survey use it for front and back ends. You also saw in the last section that students were 11% of the respondents. My guess is that there are a lot of developers coming out of full-stack coding bootcamps with JavaScript/Node.js as their primary language.

So I'm not sure this survey represents the industry as a whole, but in its defense I will say that 50% of the respondents have over six years of experience, so JavaScript and Node.js aren't just becoming the main language for new developers.

PHP took a hit

The most popular languages (measured by the number of tags on Stack Overflow) stayed fairly static over the past year. The only noticeable changes were 4% increases for Node.js and AngularJS, and a 4% decrease for PHP.  The only shift in rankings was Node.js and Angular leap-frogging the C language.

Some other interesting stats in tooling:

  • Visual Studio and Notepad++ tied for the top dev environments (34%)
  • Windows is the most popular OS (52%) when you combine all the versions

Visual Basic is the most dreaded language, WordPress the most dreaded tech

What does "dreaded" mean in terms of this survey? It's the "% of developers who are developing with the language or tech but have not expressed interest in continuing to do so." Visual Basic is dreaded by 79.5% of developers while WordPress is close behind with 74%. Matlab, Sharepoint, and CoffeeScript weren't that far behind. All three were still in the 70s.

Unfortunately for these developers, Wordpress still drives a huge part of the web.

Rust was the most loved, Android and Node.js the most wanted

It's kind of odd that an emerging, still-obscure language like Rust took the mantle of "most loved," which is the opposite of SO's definition for dreaded—"loved" means people using the language want to keep using it. It probably won because there aren't enough people using it yet to foster an ample amount of haters (along with the litany of "Rust Considered Harmful" posts that will follow).

With F#, Scala, and Go ranked just behind Rust, this kind of proves my point. The creators at Mozilla have a lot of ambition for sure, trying to build Rust as a modern systems programming language that can replace a lot of C/C++ development. The sloth's pace of Rust's development (which is not as slow as a snail's pace) tells us that the creators definitely don't want to mess this up.

Android, Node.js, AngularJS, Python, and JavaScript formed the upper tier of most wanted technologies, which means that they were interesting to developers who weren't using them yet. Developers are curious about mobile, JS full-stack, and Python if they haven't tried them.

Want to make more money? Learn F#, Scala, Spark, or Cassandra

F# keeps winning the title of being the most lucrative language to know. And every year I explain this by saying there’s scarcity (and more money, as a result) in obscurity. COBOL also pays really well, but do you really want to crush your soul like that?

Spark and Cassandra are a big deal though. Even though Hadoop has high salaries too, Spark is what you really want to learn these days.

Be careful about taking this survey data too literally. Make sure you read about the right way to use salary data when picking languages and frameworks.

React is the new hotness and Windows Phone is dead

If you follow the JavaScript community at all, you know that React is gaining momentum and possibly on its way to overtaking AngularJS. Thanks to its simplicity (and its Facebook origins), React more than tripled the number of upvotes on its questions. The second-highest trending technology in SO votes was Spark, which had a 163% increase in votes this past year.

In other news, Windows Phone is dead. This 65% downward trend could have been predicted since Microsoft has already decided to make less handsets this year. Haskell (-40%), CoffeeScript (-37%), and Dart (-27%) aren't doing so well either. Sorry Haskell hipsters!

Big surprise! Your boss makes more than you.

Among developers with 5+ years of experience, executives make the most, averaging $150K a year. Engineering managers are next with $143K. Enterprise level services developers ($122K) and mobile developers ($115K) were the next runners up. 

South African developers will eat well, Montreal devs will rent for cheap

Had a big laugh on the big-mac-index on the @stackoverflow survey results :) https://t.co/AXHhql44qw

— Stephen Reindl (@stever04) March 17, 2016

Maybe this is more of a commentary on the worldwide prices of beef, but Stack Overflow decided to measure how well developers can eat on their average salary for each country by measuring it against the number of Big Mac's they could buy. South Africa came out on top with 25,713 Big Macs and the US came in 2nd with 21,530. Too bad this tells vegetarians nothing.

For rent costs vs. developer salaries, Montreal, Canada is the winner with 13% of annual salaries going to rent. This was calculated using an average apartment rental cost in the city center and the average developer salary of the area. Berlin, Pune, Bangalore, Austin, and Seattle are not bad either.

Both of these stats used mean and median salaries only from developers with 5+ years of experience.

Push it real good!

77% of developers who commit code multiple times a day are satisfied with their jobs. That percentage decreases as developers commit code less often. The majority of developers (57%) do check in code multiple times a day.

Does this mean we should have a separate repository for questionable commits, so developers can still feel good about themselves? Continuous delivery = continuous happiness!

Diversity among developers still sucks

Uhhh.... I really thought we had a bigger piece of the pie now... @girlswhocode https://t.co/maTpwI4V7k pic.twitter.com/a6eTf19OJ0

— Cher Stewart (@cherlequin) March 17, 2016

Stack Overflow agrees that things are bad:

There’s no way to sugarcoat it. Lots more men are writing code than women. This year’s overall gender breakdown closely matches last year’s results.

There's also clear segmentation going on, because designers (13%), QA (12%), and front-end developers (11%) are the most common roles when you look at only the female respondents. Remember, the section above showed that full-stack and back-end developers were the most common roles overall.

5.9% of the 5.8% of stackoverflow users who are women are iOS devs https://t.co/HRCtTymXGx

— Tamar (@tamarshmallows) March 17, 2016

At least there's no pay gap between men and women below the age of 30. 59% of respondents were under 30, so maybe a new generation is changing things.

Most of the work to fix this problem lies in the hands of educators, parents, peers, and industry leaders, but if anyone is looking for women's success stories and tips for advancement, there are plenty of those.

76% use Stack Overflow to get help for their job

Which means that when SO goes down, so does a decent-sized chunk of our software industry.

Image credit: Jan Griffier

Call me "developer"

Among the occupational names used for people who write software, "developer" (72%) was more preferable than "programmer" (60%) or "engineer" (42%).

Rockstars and Ninjas

Stack Overflow is not without a sense of humor. They asked developers if they refer to themselves as "rocksstars" and "ninjas" often. Spoiler alert: they don't. I'll leave you with a quote from the SO analysts:

10% of respondents self-identified as Ninjas. Real ninjas don’t tell you they’re ninjas. They just sneak up on you and slit your throat, which generally constitutes "hostile workplace environment."

What was your take on the survey results?

Topics: App DevDevOps