# What’s Hot (and Not) in Tech Skills

At Dice, we have a lot of discussions about which technology skills and titles are hot and in demand right now, and which are less desirable. Such talks are important, as they drive many aspects of our business, from the number of applicants for a given job, to the salary offered by employers.

To better understand this concept, we can think of this as a problem of supply and demand. In this case, ‘supply’ is the number of job seekers who have a given skill, while ‘demand’ is the number of jobs listing a particular skill. If we take these numbers and normalize them so their values lie in the same range, then divide the demand value by the supply for a given skill, we get a number that indicates how in-demand or ‘hot’ a particular skill is.

‘Hot’ skills get a number higher than 1 (more demand than supply), while skills that are less in demand getting a number less than 1. I call this concept the ‘Demand Ratio.’

In the chart above, I have performed a supply-and-demand analysis using the top 700 technology skills and top 400 job titles in our current dataset. However, to optimize the experience for slower machines and mobile browsers, I have defaulted to 200 titles and skills; click “Larger Dataset” if you want the more complete list (that will take longer to render). If a skill or title is not on the chart when “Larger Dataset’ is selected, it is because it is not common enough right now to make the list. I could have included even more data points, but that would have caused performance problems.

### How to Use This Chart

The red points on the chart are the ‘hot’ skills, while the blue points are the less-in-demand skills. The black line represents the dividing line between ‘hot’ and ‘not’, based on our dataset. You can click on a skill or title to center it, and click once more to zoom in on that item. By default, the chart uses a log scale; otherwise the dots are too close together to easily discern due to the skewed distribution. You can switch this off by un-checking the ‘log scale’ check box. If you want to find a particular skill or title on the chart, use the drop-down to locate it. ‘Reset’ resets the chart’s current zoom level.

The chart shows some interesting trends. The in-demand skills and titles (the red ones) seem to be largely focused around DevOps work, front-end development (ASP.NET, Angular.js) and Big Data (MongoDB, Hadoop, NoSQL, AWS). The less in-demand skills include some legacy technology (COBOL, Mainframe, Solaris), as well as very common high-supply skills (Windows 2000, Visual Studio). Adobe Flash seems to be on the way out (0.33 demand ratio), while HTML5 is highly in demand (2.82 demand ratio).

If we focus on job titles, we not only see many of the same trends present when analyzing skills, but also that management titles (e.g., IT manager, Director, Marketing Manager) tend to be high supply and low demand, as presumably more people want to move into management than there are positions available. This is also seems true for project managers and business analysts, two of our most common titles, as these reflect more senior positions that often lead into management positions (in my experience).

To construct this chart I used the excellent C3.js charting library built on top of the D3.js data visualization library from Mike Bostock.

Simon Hughes is the chief data scientist of the Dice Data Science Team.

Image Credit: Dice

14 Responses to “What’s Hot (and Not) in Tech Skills”

February 03, 2016 at 12:16 am, Mike Rains said:

Nice Visual Analysis. Suggestion: put the color-code in the drop-down next to the technology name.

February 03, 2016 at 6:41 am, Jonathan Shields said:

Hi

nice idea and cool graphic. A next step could be separate charts for programming languages, platforms, databases and industry sectors. Time to Learn Java then 😉

Jonathan

February 03, 2016 at 7:16 am, Daniel Barlow said:

Interesting insight. I was surprised to see R in there but not Ruby. Do you have data on Ruby?

February 03, 2016 at 7:21 am, Nawar Noori said:

No individual functional programming languages or just ‘functional programming’ as a whole? That’s a shame, especially considering the rise in popularity of languages such as Haskell, Erlang and Clojure. Would have been interesting to see some stats on that.

February 03, 2016 at 8:16 am, Ash said:

This article does not make any sense. Visual Studio is not a skill, neither is Apple. AngularJS is a skill, which is missing from the list entirely. The title of this article should have been, “The temperature of tech terms/words.”

February 03, 2016 at 12:32 pm, PSAHA said:

Nice to see Java shining once again!

February 04, 2016 at 9:12 am, Bryan said:

Very interesting – thanks. IT management by skills has a ratio of 3.44, so I wonder what jobs are pushing that up, since manager and director titles seem to contradict it with very low ratios.

February 04, 2016 at 11:00 am, BobZ said:

I see nothing about power conversion or power electronics which as my field seems to have cooled recently. PCB layout (Using Cadence Allegro) seems to be very much in demand and I see nothing about it. I guess we cannot expect people in the recruiting field to have a good understanding of all the specialties they recruit for.

February 04, 2016 at 6:32 pm, Seth said:

The data on specific skills *might* be valid, but there’s also a lot of noise from terms that are so general most people wouldn’t bother listing them as skills, and I question the accuracy of the data on skills not directly tech-related (this probably isn’t a good dataset to use for them). Right away, I notice that the hottest skill is software engineering, which is almost 9 times as hot as software development, and that’s a huge red flag that reads, “THIS CHART WAS CONSTRUCTED WITH POOR METHODOLOGY AND SHOULD NOT BE TAKEN TOO SERIOUSLY.”

February 09, 2016 at 1:43 pm, Simon Hughes said:

Please note that this chart shows the most common skills only, as listed by people on their resumes, as mentioned in the article. Some skills, like ‘Ruby’ are listed as part of the ‘Larger Dataset’, if you check that box. However, due to performance constraints, we can’t list every possible skill or title on this visualization.

Also note that while we do some basic normalization, we have been careful not to introduce bias into the dataset by equating certain terms and phrases, or imposing our own thoughts about what is and is not a skill. Items mentioned as skills and titles in this analysis are those self-reported by technology professionals on our site in the sections marked “skills” and “current or desired position” in our profile pages. So they reflect what our users view as skills and titles, not our own opinions as to what constitutes either. In our opinion, this is the fairest and most objective and thus unbiased way of presenting this dataset.

February 11, 2016 at 3:37 pm, Hal Hall said:

Re. “The less in-demand skills include some legacy technology (COBOL, Mainframe, Solaris)….”

Remember when Sun Microsystems’ CEO, Scott McNealy, kept saying, “The mainframe is dead?” How interesting to see Sun”s Solaris lumped into the “less in-demand” category along with the mainframe and COBOL!

August 23, 2016 at 3:11 pm, Jose A Bergiste said:

This graph is great but could use some data cleaning/normalization. For instance .net Developer and Dot Net Developer should be the same item. Another example is Director of It vs Director IT. Doing this would better categorize the data.

October 05, 2016 at 11:01 am, Mark Dalsaso said:

Simon, this is a wonderfully useful example of visual analytics…via logarithmic scatter charting! If the data is from February, when will it be refreshed?