Dice Data: How Tech Skills Connect

I wanted to follow up on this post by Simon Hughes, our chief data scientist, and share an experimental visualization we created from Simon’s work. It offers a graphical way to explore the relationships amongst skills. Let me first describe how to use it, and then what went into building it.

Every circle (or node) of the visualization represents a skill. Hovering over a node will reveal the skill and its associations. Colors designate different “communities” that coalesce around skills; for example, the sky-blue cluster (bottom left) is mostly composed of skills related to customer/tech support, whereas the light green group (top right) includes “Big Data” skills:

SkillsVisCombinedShadow

Try clicking “Java”, for example, and notice how many other skills accompany it (a high-degree node, as graph theory would call it). As a popular skill, it appears to be present in many communities: Big Data, Oracle Database, System Administration, Automation/Testing, and (of course) Web and Software Development.

Visualization 2

You may or may not agree with some relationships, but keep in mind, it was all generated in an automatic way by computer code, untouched by a human.

For those interested in how we built this visualization, it involved multiple steps. We started with Gephi, an open-source network analysis and visualization software package, by importing a pair-wise comma-separated list of skills and their similarity scores (as Simon described in his article) and running a number of analyses: Force Atlas layout to draw a force-directed graph, Avg. Path Length to calculate the Betweenness Centrality that determines the size of a node, and finally Modularity to detect communities of skills (again, color-coded in the visualization).

Once the graph was the way we wanted, we exported it as an XML graph file (GEXF) and converted it to JSON format, with two sets of elements: Nodes and links. It was then a matter of leveraging D3 to visualize the data as a network.

We would love to hear your feedback and questions.

Image Credit: Dice

Comments

11 Responses to “Dice Data: How Tech Skills Connect”

July 02, 2015 at 2:05 am, Dan said:

Where is PHP? It’s absence makes me doubt the value of the chart as a whole – it’s a bigger technology than many of the others listed, yet doesn’t get a mention. How did you generate the list of items? I did read the linked preceding article, but it doesn’t say, and that’s key to the veracity of this chart.

Reply

July 02, 2015 at 3:54 am, Luca said:

Hi

Very nice idea and well implemented, but it seems very difficult to keep the content correct: e.g. why are there no vertices from .NET / C# to any database (SQL or NoSQL)?

Kind regards,
Luca

Reply

July 02, 2015 at 10:43 am, Bob said:

This is all fun, this office IT skills, but I cannot locate any skills related to “Industrial automation” in this picture.

Industrial automation is always there hidden in the background making sure everything works smoothly and safely.

For example: Electricity generation (power plants), electricity transport, Potable water, Waste water, Food & Beverage, Oil & Gas, Data Centers, (harbour) cranes, AVGs, Trains, Subways, Tunnels, Bridges, Roads, Pipelines, Ships, Amusement parks, Building automation, Elevators (the list goes on forever). They all rely heavily on Industrial automation equipment running custom software created and maintained by engineers.

So where are the related skills ?
(start with SCADA, PLC)

It must still be a mystery in the ‘Office IT’ landscape displayed here what makes our society run 🙂

Cheers,

Bob

Reply

July 02, 2015 at 12:00 pm, Mark said:

So…how does that work for someone who is an Electrical Engineering Technician? 75% of the time I deal with hardware. How would learning Python help my career?

Reply

July 02, 2015 at 12:23 pm, carol said:

Very interesting research! I would love to see the details of how this graph is visualized from those thousands of job postings. Thanks!

Reply

July 02, 2015 at 1:41 pm, Yuri Bykov said:

Mark, we had to keep the visualization to a small subset of all possible skills, otherwise it would be too busy and slow, that’s why you don’t see many EE related skills such as CAD, Verilog, PCB, Cadence, etc. As to your question how learning Python could help your career, it really depends on whether you want to stay in your field or try something different. According to our data, Python does appear in some domains adjacent to Electrical Engineering. We may publish another article focusing on career paths, which could provide better insights to questions like yours. Thanks for your feedback, Mark.

Reply

July 02, 2015 at 2:13 pm, CPhxLA said:

amazing! i’m getting a thumb graft just so i could give this more thumbs up!

how do we get this integrated into HR and recruiting? there’s so much overlap and related fields, this really could be useful in getting more understanding of the relationships within tech.

Reply

July 02, 2015 at 7:25 pm, Simon Hughes said:

As Yuri mentioned, there are too many skills (2-3k) to fit on one page, so we took a seed set of top skills and expanded the graph with skills related to those. So some will be missing. However, the dice skills pages are built on the same technology:

http://www.dice.com/skills

This has a much broader set of skills, including most if not all of those mentioned above, and allows you to see skills directly related to those skills. For example, here’s PHP:

http://www.dice.com/skills/PHP.html

Reply

July 03, 2015 at 1:04 am, Yuri Bykov said:

Will try to reply to multiple posts here:

@Dan – We have a few thousand unique common skills in the database, which is far too many to show on a single visualization like that one. So we picked a seed set of skills that were diverse but popular and then took any skills that were also common and related to those. PHP and Angular were not on that list unfortunately, despite being relatively popular skills. Regardless, the point of the post is to explain the process of discovering relationships between entities, not how comprehensive the list of those entities is.

@Luca – If you look at .NET or C#, you will have noticed links to database technologies such as MS SQL Server, Oracle, PL/SQL

@Bob – see my response to Dan above.

@CPhxLA – “graft thumb”? Funny. On a serious note, yes, HR would greatly benefit from such integration.

@All – thanks for all your comments and feedback! For more skills and interesting information, please check out our Skills Center: https://www.dice.com/skills/?icid=dicenewsES

Reply

July 03, 2015 at 6:50 pm, Graham Toal said:

GIGO

Reply

July 06, 2015 at 4:50 am, DBaldwin said:

Cool idea… but crashed my browser 3 times when I tried to use it.

Reply

Post a Comment

Your email address will not be published.