Digging Deeper Into Skills

A few months ago, in our last major announcement, we introduced our new "V2 Infrastructure," which brought with it backwards-compatible new endpoints, improved performance and data handling, and easier-to-use request structures. At the time, we also teased that we were excited about some new endpoints and enhancements to come. Well today is the first unveiling of a few of those new features!

From the beginning, when we showed off the Competency Analysis at SXSW, we've wanted to dig deeper, to go beyond the bulk profile of a piece of text and identify relevant work skills and activities on a sentence-by-sentence level. Our belief is that the combination of holistic text analysis and more detailed, sentence-level analysis provides a powerful ensemble of tools that our clients can leverage to do amazing things with work-related, unstructured text.

Today, we are very proud to announce our very first set of API endpoints that perform sentence-level skills analysis and matching. Simply put, we can now take a skills-related sentence, or a collection of sentences, and align to work activities from the SkillsEngine core library. This is an instrumental piece in our promise to help you align unstructured work-related text with a common skills language.

/v2/skills/match and /v2/skills/multi_match
Our skills/match and skills/multi_match endpoints allow you to provide us with a sentence (or multiple) and a set of contextualizing SOC codes. From there, we analyze your text, take it apart, and feed it into our matching algorithms, which search our database of over 4,000 detailed work activities looking for skills that may be highly related or relevant to the text provided (and contextualized by the occupations you optionally specified).

Combine this new endpoint with our Competency Analysis endpoints and you have a powerful two-step tool. First, use the Competency Analysis to generate a holistic profile of your entire text; we'll give you back related occupations as SOC codes, as well as all the other library elements you've come to expect. Then, send our skill matching endpoint your skill-related sentences -- learning outcomes, job duties, "things I've done" from a resume, etc. -- and the previously-determined SOC codes. With that data, we can provide you a list of specific, detailed work activities that are directly relevant to those sentences in your document.

Beyond the new skill-matching endpoints, our "v2.1" release also includes several other enhancements, changes, and new utility endpoints:

A new version of the competency analysis that provides General, Intermediate, and Detailed Work Activities in three separately-scored lists, instead of a hierarchical tree. This has been a hot request from a lot of our API users, and we're excited to finally roll it out.

Element Exclusion
The full competency analysis can be a beast, and not every consumer of our API uses everything it provides. A new optional parameter is now available for both formats of the competency analysis endpoint that allows you to specify a list of library elements that you do not want us to return. Feel free to use this list to pare down the response from the competency analysis so that it only includes the specific items you're using. And it's not just a convenience -- excluding elements can actually make the response faster, since we don't have to do the work to round up that unneeded data!

A number of users have asked for an endpoint that simply returns the entire set of 1,110 occupations in our core library. To keep things fast and simple, this endpoint does not include labor market data. It's just a quick way to get the basics about all of the occupations we track, including their SOC codes.

Similar to the occupations list, we're happy to provide a new endpoint that lets you look up specific occupations, by SOC code. This endpoint does return labor market data, and you can provide an optional state_code to get data specific to your favorite U.S. state.

You can check out all of the new endpoints and features in more technical detail in our API Documentation.

With all of those changes, enhancements, and new features, it might seem like we're way down the road, but we're just getting started. Keep an eye out for lots of great new things to come, including what we believe will be the most robust military-to-civilian occupation translation endpoint ever made available, ever-ongoing curation and library improvements, feature and performance enhancements to all of our existing endpoints, and our first user-facing web application, just over the horizon.

As always, we are grateful for our amazing API partners and excited to see what kinds of cool new things you do with the features we've added to SkillsEngine. If you have any questions, comments, or just want to show off something you've used our systems for, please don't hesitate to contact us. And if you have technical issues connecting to or using the API, we're always hanging around in the Support Forums, eager to help.