Select Page

Tailored User Profiles for Navigable Careers: Ontologies for Knowledge Management in the JobTeaser Suite of Psychometric Tools

Data Integration & Ontologies

Our customer

Operating primarily out of France with a network that spans into other European countries, JobTeaser was founded in 2008 with the goal to give career guidance and recruiting opportunities to young people. Since 2018, JobTeaser’s outreach has expanded to include a rich ecosystem of psychometric tools that focus on preparing students who are currently enrolled/recently graduated from university with comprehensive overviews of their career strengths.

The challenge

In order to match users to the most accurate profiles, the JobTeaser platform assesses many aspects ranging from psychology, salary expectations, skill profiles, job definitions, etc. With such a diverse baseline, JobTeaser requires a sophisticated semantic approach that integrates multiple facets of data, defines concepts, and makes ontological relations.

If finding a job were as easy as swiping right for a date on Tinder, people would have a better time hunting for jobs and submitting applications. Luckily, a solution lies in the friendly job application platform called JobTeaser, whose suite of psychometric tools helps match users to job profiles and professional skills, in hopes of leading them down career paths that best suit them.

JobTeaser’s mission is simple: to help students find their way career-wise, by helping them navigate the difficult job market and ultimately understand who they are as workers, what they want in a career, and how they can obtain them. 

To reach this mission, the PoolParty Semantic Suite’s Thesaurus Server is their software solution of choice, which was sought out by JobTeaser’s Product Manager and Senior Information Architect François Violette to ensure that the JobTeaser suite of tools could be enhanced by controlled vocabularies and ontology knowledge management.

Exploring careers with JobTeaser.

JobTeaser comprises a suite of tools dedicated to matching students to jobs that fit their user profiles. Currently, each of the tools are available in French, as that is the primary language of the company. However, some of the tools are available in English and German, with current progress being made to ensure that they can all be available in these languages in the future. PoolParty’s multilingual capabilities help this adaptation. 

Having kept the theme of the career journey in mind, each of the tools is aptly named from a famous explorer.

  • Marco is a questionnaire that assesses the students’ professional interests. 
  • Vasco, otherwise known as the “Business Tinder” according to the JobTeaser team, matches users to the job roles for students to explore, based on their profiles. 
  • Freya is a tool that measures the students’ occupational values.
  • The Hester questionnaire asks students about their soft skills and generates a report with the results.

Using linked open data integration as a vehicle for JobTeaser’s diverse profiles. 

 

Altogether, these tools give the JobTeaser users comprehensive profiles of their career aptitudes, going beyond the traditional CV and cover letter expected of applicants. The tools are built from a number of factors ranging from psychological theories for personality tests, external research for job markets, science for modelling, analytics for measuring, and the list goes on.

It is important to note how the tools are made from this diverse framework, because it is semantics that helps to incorporate and organize the various factors into a powerful, working platform.

Thanks to PoolParty’s linked open data capabilities, the external research and data that JobTeaser’s teams use to build the tool can be easily integrated with their internal terminology and metrics. 

According to Violette, “there would be no way to integrate the information into our products [without PoolParty] because all the data points we measure for students are outside the [JobTeaser] ecosystem.”

Photo shows diagram of JobTeaser's data architecture.

Photo courtesy of JobTeaser, via PoolParty Summit 2021.

The JobTeaser team has their own taxonomy for skills and roles, but they also rely on ESCO, the official multilingual taxonomy of common skills, qualifications, and occupations in Europe, to supplement the data the team has already developed. The integrated ESCO terms are matched against the internal taxonomies in PoolParty so that the JobTeaser tools can continuously reference both sources of data. 

As noted in the screenshot from Violette’s slide deck at the PoolParty Summit 2021, PoolParty acts as the “orchestrator and source of truth” where the concepts and vocabulary built from various data references can be recalled for scoring metrics in the JobTeaser applications. 

Defining concepts for a strong backbone.

 

In the PoolParty application environment, each role is defined as a concept. Simply put, a concept is a unique way to “package” a term so that it and all its surrounding details can be bundled together in one place. Technically speaking, a job role defined as a concept helps to label and organize different components about the job role as well as classify it into a taxonomy that models its hierarchical relations.

Speaking in terms of performance, using concepts means that the JobTeaser tools can test and detect the little nuances between professional dimensions, career domains, skills, roles, etc.

If we define the job role Social Media Specialist as a concept, we can discern the following information as pictured in the employee ID illustration.

Soft skills, hard skills, professional domains, and career domains all make up the concept. In PoolParty, additional notes can be added to the concept regarding role definitions, photos, and so on.

All these details are maintained in one place so that the JobTeaser tests can easily determine user profiles based on the various skills bundled in the concept. A user who displays high scores in these various traits suggests having a tendency towards the social media specialist role.

This infographic is an example of what a Social Media Specialist role would look like when defined as a concept.

JobTeaser’s controlled vocabulary and ontological skeleton.

 

As we have seen, concepts in a taxonomy provide a method for organizing the data. They are the backbone to what Violette refers to as the “skeleton to help project the different links between them” in an ontology. 

Since the JobTeaser platform uses a lot of different information in order to provide results, it’s essential that the information is maintained centrally. JobTeaser uses taxonomies as the vehicle to this approach, with additional strength lying in the fact that the taxonomy can help them to standardize their terminology enough to make it actionable.

Prior to PoolParty, the product and data teams that were modelling the domains struggled to make connections between concepts without a controlled vocabulary. A controlled vocabulary refers to a standardized method of naming concepts so that they are organized in a system of sorts. 

Since all the details of a particular job role were defined in one concept (as pictured in the ID card), job roles could then be indexed properly and more easily retrieved by the JobTeaser testing tools. The controlled vocabulary ensures that the job role, skill, professional dimension, etc. could be referred to by the data scientists and the JobTeaser team using the same preferred term instead of all the individual details defined under it. 

In other words, universal naming conventions allowed the people working with the data to rule out any ambiguity between concepts and define how they are linked to other concepts. With the controlled vocabulary, Violette and his team were able to map out ontologies to add a layer of semantic expressivity.

This diagram shows how the job role and skills are mapped in an ontology

Using an ontology like the one shown in this example, grants the ability to make links between concepts and thus provides a framework for the JobTeaser’s tools assessments. The ontology maps the top level classes (the circles) with its individual attributes (ovals) and marks the relations (arrows) between each class. It is with this framework that the tools can determine that a person would be good at social media management because they have entered that they have strong video editing and social perception skills in the test, which are requirements for the social media specialist role.

At the same time, high scores in these specific traits mean that the user will not be an information architect for example, because as Violette puts it, the details embedded in the concepts determine “what it is, and what it is not.” In this way, the tools are smart enough to avoid overlap so that they can give accurate profiles and recommend jobs that are actually relevant and useful. 

The behavior is similar in all of the JobTeaser tools regardless of the specific dimension that is being assessed in each of the tests.

The Vasco tool, aka the job role Tinder, presents a new job role with each swipe right by the user. The tool presents various roles with various underlying skills until the tool’s “algorithm” has understood the user’s preferences and profile so there are fewer variances between roles that display.

Photo shows how the JobTeaser Vasco tool responds to user input to recommend tailored roles.

Photo courtesy of JobTeaser, via PoolParty Summit 2021.

 In other words, the more a student swipes right on roles that lean towards analytic roles, the less they will be presented with writing roles. Once this is the case, students can review their job role selection, check job offers from JobTeaser for that role, or talk to alumni within their university.

In the Marco tool that analyzes professional interests, a person scoring “Innovator” in the first position and an “Organizer” in the second position, does not have the same profile if the positions of these same interests are reversed, even though they are still the top two most prominent interests. The semantics in PoolParty can analyze these subtle nuances with concepts and their relations in the ontology.

The PoolParty UI and its interoperable approach.

Considering all the data that had to be integrated and managed by multiple people from the taxonomists to the data scientists to the web agency, interoperability and a friendly user interface were important aspects of choosing PoolParty as their software. 

As a result of their vast ecosystem, it was imperative for Violette that the semantic software was accessible and “quick to ramp up.” PoolParty has allowed easy onboarding of employees to data management tasks, and has helped them quickly distribute the data in a clean, understandable fashion to the web agency that develops the front-end JobTeaser tools. 

Having worked with PoolParty in the past, Violette was keen to use it again for its ability to act as a “centerpiece that all things depend on.” With PoolParty as this centerpiece, he and his team could benefit from many things that go beyond just cleaning up the database. In the future, the team wishes to expand upon the controlled vocabularies and ontologies to create rich knowledge graphs of data.

As Violette puts it “companies who realize that middleware is a gamechanger go miles with it.”

The Story at a Glance

Want to learn more about JobTeaser? Head over to their website for more information.

JobTeaser was one of our honorary use cases at the first ever PoolParty Summit 2021. Click to hear more about their use case and watch Violette’s extended presentation for free.

Concept tagging icon

Download our Case Study Booklet: Business Solutions with Semantic AI