Examining the Hottest (and Coldest) Tech Skills
Last year, Dice launched an interactive visualization, created by Simon Hughes, that broke down supply and demand for various tech roles. In this context, “supply” is the proportion of job seekers on Dice with a particular skill, while “demand” is the proportion of jobs listing that skill. We took those values and normalized them so that they sat on the same range; by dividing the demand value by the supply for a particular skill, we get a number that indicates how in-demand (or “hot”) a particular skill is. This year, we’ve updated that visualization, incorporating some of the broader feedback we’ve received over the past 12 months. We added data, considered different modes of presentation, and sped up label generation—just move your pointer over a skill dot, and see its name appear. We doubled the number of available skills (to 1,400), and now enable users to pull out specific categories of skills such as programming languages (the hottest and coldest are automatically labeled). We’ve also introduced an alternative chart type, known as the “bee swarm” plot (more about bee swarm), which attempts to plot all skill data-points along a single “heat” axis. This lets you quickly discover the hottest and coldest skills. You can also split individual skills into their demand and supply parts. [caption id="attachment_141354" align="aligncenter" width="445"] "Hot" skills.[/caption] For example, by pulling out the ‘Software Development’ segment, we can quickly see that Java-based web enterprise skills such JavaBeans, Java Servlets, and JNDI are relatively “cold,” while open source, cloud deployment (Puppet, Chef, DevOps) and front-end frameworks (Ruby on Rails, Knockout.js, Backbone.js) continue to stay “hot.” This is something we'd like to maintain and extend going forward, so please let us know how you’d like to see the visualization evolve. We'd also like to re-include job titles once we settle on the best way to incorporate them. This interactive visualization was developed by Si Hammond using the excellent D3 framework, version 4.