The biggest takeaways from StackOverflow’s Developer Survey 2018

AI, Ethics, Programming languages, Health, Diversity and Salaries Yesterday released from their annual Developer Survey. StackOverflow this year’s data Their recap is always interesting and full of trends and helpful stats, but it’s also quite long. Not all of us have the time to sift through 60+ pages of bar charts and percentages. Fortunately, I do. I know I’m a dev and should go digital, but I still do love to just use a pen & some paper Having carefully read over it a few times, here are the main things to take away from this year’s stats. Let’s start with a confirmation of an intuitively known fact: Code — Developers love it. This one makes a lot of sense but a reminder never hurts. Developers are not (only) in it to make a living, they seem to truly enjoy as regardless of their job status. programming over 80% of developers code as a hobby Actually, to devote to it (because they exercise daily, spend more time outside or have care-taking responsibilities) seem to be those who code the most. the developers with the less time To add to that**, every other developer also contributes to open-source software** (OSS), although some languages seem to be especially favored by OSS contributors, such as Rust or Clojure, with 70% of its developers being OSS contributors. To confirm this trend even more, , and . Which is coherent with the fact that . over 3/4th of all developers say they go to Hackathons is simply that they enjoy them half say they go to them to improve their programming skills 87% developers have learned a technology or programming language by themselves Point is, we love to code, and we code a lot. Health — Bad habits, good furniture. Our love for code might have an impact on our health though, as . 1 out of 3 developers will skip at least a meal per week to be more productive But despite our bad habits, and 1 out of 2 developers uses a standing desk just as much have an ergonomic chair, ergonomic mouse. Tech — What’s in fashion? That’s where the money is after all (although you might say that’s more applicable to the upcoming section). So let’s get to it. Salaries Honorable mentions Rust sails through its 3rd year in a row as Most Loved Programming Language Visual Basic 6 (VB6) through its 3rd year in Most Dreaded Programming Language JavaScript is here to stay — and we seem happy about that. , it’s also the . JavaScript is the most used Programming Language once again 2nd language that developers want to try the most and people seem to be right to want so, since NodeJS, React and Angular (the 3 most used frameworks across all languages) place 1st, 2nd and 4th respectively as frameworks that developers want to try the most they place 2nd, 4th and 7th on the list of the most loved frameworks. a regular part of most JS stacks, is MongoDB, the database most developers want to try. Redis & PostgreSQL are the darlings of the database world. MySQL and SQL Server are the most used databases, with PostgreSQL coming up as third. But , and PostgreSQL comes up as 2nd most loved database technology 3rd most wanted to try. It’s in good company**,** as and . Redis takes 1st place for most loved for the second year in a row 4th as most wanted to try Python still rises After dethroning PHP of its #8 spot last year, Python appears to be the fastest-growing major programming language. Python now takes over C# for the spot of 7th most used language. For the second year in a row, . It’s also the Python is the language most developers want to try 3rd most loved. Its Machine- framework, Learning PyTorch, is the 3rd most loved framework. TensorFlow dominates the Machine Learning world And not without reasons since for the respondents. TensorFlow is the 3rd framework most developers want to try. it’s the top, most loved framework AI — Overall, we’re excited. Concerning AI, it appears that as developers we’re more excited than afraid. The survey studied four aspects of AI: The automation of jobs AIs becoming smarter than humans (the singularity) The definition of “fairness” evolving in human/algorithmic decisions Algorithms making important decisions And as it turns out But as could have been expected from the masters of automation we don’t seem to find one prospect to be scarier than any other. we’re mainly excited about the automation of jobs over the others (40% of developers find it exciting). We also mostly tend to think Only 1 developer out of 4 thinks that it should be a governmental/regulatory body’s responsibility. that studying the implications of AI and its ramifications is a task that should befall on the people designing and building it. Ethics — We have a moral code, but it still depends. 95% of developers would hesitate before writing clearly unethical code. 58% would flat-out refuse. 95% of developers would report it depending on the situation. 13% would do so publicly. This is a good news for our moral integrity, with , but within the remaining 95%, , no hard and fast rules in ethics for developers but only 1 developer out of 20 being ready to go against what they think ethical about 35% would adjust their behavior to the situation 80% of us think that we have an obligation to consider the ethical implications of our code. Ultimately though, most of us (57%) think that management is responsible for them. Diversity — Coming up! The next generation of developers still won’t be quite the melting pot, but we’re getting closer. 1 out of 3 programming students are people of color, where only 1 out of 4 professional developers are. The increase is especially noticeable for students of Middle-Eastern (148% growth) and East Asian (145%) descent. Sexual orientation and gender also see some diversification when looking at students and developers with little experience: Where 6.6% of developers identify as Gay, Lesbian, Bi or Queer in the professional world, 8.3% of student developers do so. The rate of women in the industry also rises from 6.6% to 7.4% between professional developers and students developers. when looking at the student population. Interestingly, these also tend to contribute 1.5 to 2 times as much to open source than their cisgender peers. The number of trans and non-binary individuals in the community is also on the rise Building a diverse company Where , . men seem to value compensation and benefits as the main criterion to pick a job minorities seem to value the company culture first More specifically: Women value the company culture, the opportunities for personal development and the technologies used about equally (16.9%, 16.8% and 16.4%) Trans and non-binary people value the company culture way more than any other criterion (5–6% lead over the second most important) Cisgender developers consider the diversity of the company to be the lowest priority (out of seven possibilities) when assessing a potential job, whereas it’s the 4th most important for both non-binary and trans people. Salaries — Follow the money We’ve all seen float many articles about how “X is language to learn in 2018 if you want to make the big bucks.”. Let’s see how those hold up next to the data: the The Top 5 languages with the highest-median salaries this year are: (in USD) programming F# — 74K OCaml — 73K Clojure — 72K Groovy — 72K Perl — 69K Rust — 69K The Top 5 positions with the highest-median salaries are: (in USD) Engineering Manager — 89K DevOps — 72K Data Scientist — 60K Business / Data Analyst — 59K Full-Stack Developer — 59K The position of Data Scientist is notable, as it is yet has . one of the highest median salaries one of the lowest average number of years of experience (60K for an average of a bit under 6 years of experience).DevOps are more experienced but are still more compensated on average than their non-DevOps peers (72K for about 8 years of experience on average). These are the statistics and takeaways I’ve found the most interesting from this year’s report. I have no doubt some other fact and stat I’ve decided not to include here might be of importance to you and I strongly advise reading through when you find the time to do so. the original report I hope my digest was enough to give you a comprehensive view of the current state of the industry as of 2018. If you’re one for data points I’d advise you to wait for a few weeks until StackOverflow publishes the whole dataset, I have no doubt more detailed articles will flourish then.