Hacker News Tech Hiring Trends | Hacker News Reader

Hacker News Tech Hiring Trends

ryan-williams.net

63 pointsmountaineer47 comments

If Hacker News is a leading indicator, interesting to see that JavaScript is becoming so dominant, while the dynamic languages continue to be in strong demand. You can view direct comparisons too.

Ruby vs Python vs Node vs PHP [1] Angular vs Ember vs Backbone [2] iOS vs Android [3]

[1] http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

[2] http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

[3] http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

I've seen this trend forming over the last 3-4 years. Backend technology now has a vast wealth of frameworks, tools and hosted services to choose from. People still bicker over Django vs Rails or even Spring, but regardless any of these are so vastly superior to what we were working with 10 years ago. Cloud hosting has been another huge disruptor. We still have plenty of work, but on a typical web project nowadays, client-side development has seen increasing demands and slower uptake of new technology thus shifting the project bottlenecks form backend to front. Expectations are higher since the capabilities are there, but we've only just passed the hump of supporting IE6 and now we have to worry about phones.

Windows Phone is not even there :(

Not due to its uptake though. This one fell through the cracks in the analysis. Multi-word phrases aren't supported yet and I didn't have that in my initial dictionary.

There seems to exist a lot of false positives; for instance, I included Lisp on the chart, and it indicated 7 job posts in May '13. I went to the whoishiring's posts for that month (both the "Who is hiring" [1] and "freelancer?" [2]) and I counted 6, but only one of them (the first instance, in a "top level" comment) is actually a job posting (all the others occur in sub-discussions). And as it turns out, it seems that it's only a sort of "eye-candy" (although I admit that filtering these may be harder).

My point is, you might be able to avoid false positives if you only consider "top-level" comments.

[1]: https://news.ycombinator.com/item?id=5637663

[2]: https://news.ycombinator.com/item?id=5637667

Yeah, I agree. I'll update to include just top-level comments.

I really like this. If I was going to suggest a change/tweak, you might want to define the chart colors based off what is in the compare boxes rather than going with Highcharts' defaults. When looking at a compare and adding additional entry(ies) you need to constantly look back to the legend.[1][2]

[1] http://www.ryan-williams.net/hacker-news-hiring-trends/?comp... [2] http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

I appreciate the feedback, are you doing multiple comparisons with one constant? I'll take look at this tweak for the next round.

The first time I noticed it was comparing 3 values (e.g. JS frameworks) and adding a 4th (e.g. JS). I would imagine this is only an issue if the 4th addition is not the least popular.

But again, this is really cool.

A higher percentage mentions of JavaScript without jQuery in 2013 than previously: http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

This could also mean that jQuery is becoming such a standard that nobody bothers to mention it explicitly. Does anyone remember PrototypeJS and mooTools?

I was really impressed at an O'Reilly conference to discover how on top of this sort of thing they are. Tim O'Reilly is huge on data and would be watching web-trends for emerging technologies like a hawk to decide what books to commission next.

Great visualization! It might be interesting to normalize mentions by the number of other languages mentioned in each comment. Something like "Looking for JavaScript hacker" is a much stronger vote for JavaScript than "Looking for someone who knows Ruby, Python, and HTML5/CSS/JavaScript. Erlang knowledge is a plus." I think that while JavaScript is growing on the server-side, it's going to be overrepresented due its dominance on the client side.

Weighting will be a challenge, but is a good idea, thanks. I'll take a look at some NLP libraries and see if this can be improved.

The YOY stats are misleading when it's not a cumulative or smoothed. Look at the Rails line, it'd be hard to argue that there has been a decrease in rails over a year but literally Aug 12 to Aug 13 has seen a "20% decrease".

Some smoothing would better show the real picture.

Good catch, the first version is a little raw. I'll be working to improve the calculations/methods, including this.

FYI Your site is blocked by my corporate firewall as a site containing "Malicious Sources".

Thanks for the heads up, what brand is the firewall? I know the site had some links injected a couple years ago, and I've been watching to see lingering effects.

Very cool! I was trying to compare Java, Javascript and C++ :

http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

It seems that or 'C++' is being parsed wrong (C%2B%2B) or is seriously unpopular.

EDIT :

I managed to get it to work by manually changing the parameters

http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

Uh-oh, I'll get that fixed up, thanks for catching that.

Update: fixed to decode url params.

C# is also having the same issue

Thanks, fixed.

Depressing.

I'll tell you why: My Boston-based, recently acquired startup can't find people at all, so seeing hiring increase in the technologies we use (Javascript, Python, Angular) means that trendier startups who pull stunts like sponsoring free beer at the Boston Python meetup (Thanks Bullhorn!) are getting attention where we can't. And that sucks for me, because I need someone to replace me now that we've been acquired, so I can move on to my next project.

We're under corporate pressure to convert to C# and .NET as a result of the Python hiring crunch. I fear that this particular cure might be worse for us than the inability to find Python people.

Interesting, I would think that there would be a lot of interest in joining a more established company (vs risk of startup) and use some of the new, shiny tools. I'm a C#/.NET by trade as well, and this would apply for me. I'd recommend sticking to your gut on the Python track rather than converting. Of course it depends on the type of apps and the team, but I'd agree with your take. Good luck with your search!

I would say it's more the uncertainty of recently acquired startup. The startup I was working previously(also Boston based) was hiring .Net guys and converting them to Python devs.

I was surprised to see a technology on the list on the up and up which I could not identify, kafka. There aren't that many posts submitted to HN on it and the few that are have almost no comments. Who's using it in production besides LinkedIn?

I was also very surprised to see that NLP had declined and that machine learning was not on the list (but Statistics was).

What were the criteria used for those included on the top ten lists?

Were the named tokens measured automatically collected or did you manually enter the tokens?

I manually created the tokens. Machine learning fell through the cracks on this first version, I'm adding multi word concepts. The top ten were those that had mentions both this year and last and saw the biggest change in either direction. I need to add some minimums though so going from 3 mentions to 2 doesn't register as "falling"

Right now there are four Go-related submissions on the HN front page. Ryan's tool shows just two 'golang' job submissions over the whole observed time span of Apr2011-Aug2013. Is this an indicator of complete absence of Go's usage in real-world commercial projects as of today (but what's up with all the submissions then?) or just a data parsing/interpretation issue (term Go is not present in the dictionary, while the golang is there)?

You're right, my analysis is coming up short on go. I have golang in the dictionary, but not just "go" since it would skew for general use of the term not referring to the language. I was curious about this one too since it's been seeing a lot of time on HN.

Actionscript 200% change ? Should be some min limit to the number of occurrences of each keyword to be included(like 1 post and then next year 3 posts would be 200% but it can't really represent a trend). Or is it that Adobe Air/Flash is really getting popular as mobile game dev platform ?

Good idea, I'll add some bounds as I make some changes going forward.

This is really cool. The last who's hiring, at around 4pm EST I did a find for "Ruby" and for "Javascript". Javascript had almost double the openings than Ruby. It's not as scientific as this, but I plan to continue to compare and check, just for my own satisfaction.

Thanks! Feel free to check out and review the code too to see if I missed anything.

Is this just an aggregation of keywords found or is there more to it? Because I didn't see anything like UI, UX, Design , Marketing, Growth hacking or any of those keywords? Surely those kind of positions are being offered/sought as well?

Good idea. I built a dictionary of terms to count/analyze, which mostly consists of technologies, frameworks, languages. It's at 135 entries now, but I can add some of these terms to it as well and re-analyze.

This is very interesting!

JavaScript sky rocketing makes the most sense as more and more companies are keen to delve into the Node.js world.

Not surprised at all to see EC2 falling, the only slight surprise would be Rails falling.

Feeling good that I spent the weekend learning Angular and Puppet.

Nice, great choices indeed if these trends continue.

Very cool. Thanks for this. Next step, predict the future of what employers will want (employers would love this too) :P

`angular 1700.00%` wow…

by the number of Go articles I see on hn I thought there would me more jobs related to it.

I need to do some work on identifying go, hard to tell the difference between "go" meaning golang and "go here to apply" or similar.

Really like this idea; great tool, Ryan!

Nice chart and very similar to our recruitement—since spring we've been hiring Node/JS devs only.

Seeing as you are in Berlin, would you mind sharing how many cool tech companies in Germany require you to speak German? Yours doesn't I know that.

Wow! What happened to Sinatra?

nothing special: it went from 3 mentions to 1 mention over the recording period (a "66.67%" decrease), all the while bouncing off the x-axis but possibly evidencing an upward trend:

http://www.ryan-williams.net/hacker-news-hiring-trends/?comp...

It could be an aberration, as one commenter noted there is no smoothing. However, in my own experience of building both Rails and Sinatra apps, I've tended to go more Rails, even for smaller stuff, as it just takes care of more things.

Reply on news.ycombinator.com