Now Live: Stack Overflow Developer Survey 2017 Results
What do developers want? And how do they get the job done? Every year since 2011, we’ve asked developers these questions (and many, many more) as part of our annual Stack Overflow Developer Survey, the largest and most comprehensive survey of developers in the world.
More than 64,000 developers told us this year how they learn and level up, which tools they’re using, and what employers should know about working with developers.
Our first analysis of these results is now live.
Check out the findings now to see which technologies are the most popular, most loved, and most dreaded in 2017. You’ll also learn how much money different types of developers make on average, which benefits developers care about most, and how developers want their performance to be evaluated. You can even find out how developers prefer to pronounce “GIF”. Spoiler: 6% enunciate each individual letter (“gee eye eff”).
Why do we do this each year? As the world’s largest and most trusted community of software developers, we run this survey and share these results to improve developers’ lives: We want to empower developers by providing them with rich information about themselves, their industry, and their peers. And we want to use this information to educate employers about who developers are and what they need.
We learn something new every time we run our survey. This year is no exception:
- A common misconception about developers is that they’ve all been programming since childhood. In fact, we see a wide range of experience levels. Among professional developers, 11.3% got their first coding jobs within a year of first learning how to program. A further 36.9% learned to program between one and four years before beginning their careers as developers.
- Only 13.1% of developers are actively looking for a job. But an additional 62.1% of developers are interested in hearing about new job opportunities. (If you’re an employer, Stack Overflow Talent can help you recruit both active and passive candidates.)
- When we asked respondents what they valued most when considering a new job, 53.3% said remote options were a top priority. A majority of developers, 63.9%, reported working remotely at least one day a month, and 11.1% say they’re full-time remote or almost all the time.
- A majority of developers said they feel at least somewhat underpaid. Developers who work in government and non-profits feel the most underpaid, while those who work in finance feel the most overpaid.
31 Comments
JavaScript is the most popular language for data scientists? Is that right?
Doubtful indeed! I reckon it’s more like this: most data scientists use predominantly another language – Python, R, Scala, Matlab or something yet more domain-specialised. They’ll pick _one_ of these and use it whenever possible.
But most of these guys also _occasionally_ need to do something (e.g. web stuff) that can’t be done in their main language, and for these tasks basically everybody needs to dabble with JavaScript, SQL, PHP. Also, most of the data languages also aren’t particularly fast but can easily call C code, thus most who are dealing with sizable amounts of data will also delegate some bottlenecks to C or C++ every now and then.
Sounds about right. Remember, the results listed under that table account for multiple responses. This just means that out of all data scientists, a larger proportion know javascript than any other language. This makes sense (at least to me), as javascript is overall a very popular language. If you think about it, what languages would you expect to be popular among data scientists? R, Python, Matlab, Maple, Mathematica, etc. (perhaps SQL?). I would expect (but don’t have the data to back it up) that there would be more segmentation among those languages, as not everyone would learn all of R, Maple, Matlab and Mathematica, i.e. the majority learn one or two of these languages well.
Making good numbers can be done in a lot of languages. Reporting those numbers usually involves JavaScript. “I saw this thing on the internet…can you do that with our data?”
“Men of White or of European descent were much less likely to agree or strongly agree that diversity is important than men of any other ethnicity.”. Disappointing but something we can all work on and improve upon in the future. https://stackoverflow.com/insights/survey/2017/?utm_source=so-owned&utm_medium=blog&utm_campaign=dev-survey-2017&utm_content=blog-link#work-who-values-diversity
Why is that disappointing? Any particular reason you think I should care what gender or skin color my colleagues have as long as they’re competent and can get the job done?
This is why you should care: https://www.scientificamerican.com/article/how-diversity-makes-us-smarter/
http://smallbusiness.chron.com/negative-effects-diversity-workplace-18443.html
Those are more of obstacles for management to overcome rather than critical issues with the concept, and the fact that it comes from the Small Business section of the Houston Chronicle as opposed to Scientific American should tell you that.
Of course this isn’t a critical and is obviously overcome (Racial discrimination is illegal in most places). But it doesn’t just bring good things, it’s always a trade off. Telling companies to hire more brown people because they are brown and then saying that it is the correct trade-off isn’t intellectually honest.
It’s not as simple as “Telling companies to hire more brown people because they are brown”, and no one here is denying that resources have to be expended to reap the benefits. Intellectual honesty is doing a minimal level of research on how to actually boost diversity in a company before engaging in a debate on the topic: https://www.google.com/search?q=how+can+i+boost+company+diversity&ie=utf-8&oe=utf-8
It’s not so much the skin color as it is a different perspective when reporting concerns or offering solutions. I think the best solutions can be generated not just by collaborating with like minds, but also collaborating with different minds.
Thus, different minds would typically result from different backgrounds and experiences that aren’t shared by “mainstream”.
Remember: diversity of opinion over diversity of appearance. If you preach for the latter, you are a racist.
Is PostgreSQL not related to any particular technology? I don’t see it on the ecosystem graph: https://stackoverflow.com/insights/survey/2017#technology-correlated-technologies
It looks like India has the 2nd-most responses, but most of the results analysis that is broken down by country only shows US/UK/Canada/France/Germany. Why is India left out?
Hi Kip – we’ll be releasing more analysis about regions in the coming weeks, and we’ll release the anonymized survey results in a few weeks for anyone to explore. Thanks!
why you didn’t analysis on the beginning? 😀
Do they hate Indian programmers :p ?
good job, we need to involve more devs in the future in this survey
“Among professional developers, 11.3% got
their first coding jobs within a year of first learning how to program.
A further 36.9% learned to program between one and four years before
beginning their careers as developers. ” I am surprised that this sentence is related to “been programming since childhood”:
Asking how much time between learning to program and earning programming is different from at which stage of the life one has started to program/write code. Maybe you don’t want to show the age at which they learned/got their first job as a programmer?
I wish I knew what these people were doing, and how; I’d love to (finally) move from hobbyist to professional.
Hi @davidrhysthomas:disqus you might be interested in blog post: http://haseebq.com/how-to-break-into-tech-job-hunting-and-interviews/
Why there is no mention of Angular(2+)?
I was sorry to see the compensation data filtered out those who earned more than $200k/yr. This artificially lowers the expectations for both employers and contractors. The explanation was that this was 1/2 of 1 percent of the respondents. But they didn’t explain the geography of those results. I know two C# / WPF engineers who earned between 200k and 250k each year for the past 5 years, in Silicon Valley. Their several employers build PCs and medical instruments.
I was sorry to see the compensation data filtered out those who earned more than $200k/yr. This artificially lowers the expectations for both employers and contractors. The explanation was that this was 1/2 of 1 percent of the respondents. But they didn’t explain the geography of those results. I know two C# / WPF engineers who earned between 200k and 250k each year for the past 5 years, in Silicon Valley. Their several employers build PCs and medical instruments.
Why are these engineers recruited back for multiple contracts with multiple employers? Because the management is buying insurance. These engineers will get the job done; help build the right product, solve the right problems, communicate status and issues clearly and quickly. They bring a professional approach to their teams. They reduce risk.
So it seems unfair to filter them out of the data used by StackOverflow fans who are managing their own careers. For the record, these engineers each have 20-30 years of experience.
Can you address why we don’t see any ASP, ASP.NET, or ASP MVC mentions in there. Seems like a very large percentage of postings want some version of ASP. Something in the area of 15% of websites use ASP so I would expect some traction there. Did results just fall off to the point that ASP options ended up in the “other” bucket?
Thanks for the survey. It is always required to know the problems and whats going in the minds of the employees. Keep sharing such nice updates.
I am really peeved at how these results are presented.
Your survey is not a true random sample, which is the only statistically sound way to do such a survey. I mean, this is Statistics 101 stuff. Your survey is open to literally anyone who clicks a link and has a StackOverflow account. This means you can generalize your results to StackOverflow’s membership AT BEST, and not at all to the developer community at large.
One only has to look at the numbers collected by the Bureau of Labor Statistics to see how far off the mark the numbers really are: https://www.bls.gov/oes/current/oes_nat.htm#15-0000
The way you interpret your survey leads us to believe that the overwhelming majority of programmers do web development, but there are far more systems programmers in the population as a whole than your survey presents.
I admire your effort, I really do, but please, for the love of all that is holy, find a real statistician and get some help with this thing.
“The way you interpret your survey leads us to believe that the overwhelming majority of programmers do web development, but there are far more systems programmers in the population as a whole than your survey presents.”
Anyone idea of the stats on that – web v system – programmers? I’d like a rough ratio.
This is pretty broad but it’s a starting point: https://www.bls.gov/oes/current/oes_nat.htm