Digital skills in Europe: the use of ontologies to improve the search for candidates

The increasing concern of European authorities for the management of digital skills among the working population, especially young people, is a growing reality whose management does not seem to have an easy solution.

 “The internet and digital technologies are transforming our world. But existing barriers online mean citizens miss out on goods and services, internet companies and start-ups have their horizons limited, and businesses and governments cannot fully benefit from digital tools.” This strong affirmation opens the European Commission (EC) web section dedicated to digital single market. The digital single market is one of the strategic priorities of the EC and, within it, one of the action lines is the development of digital skills among the employed population of Europe - especially young people.

When classifying digital skills, we have the same problem as classifying emerging technologies. Most efforts in this area focus on the establishment of a hierarchical classification of skills / technologies, which rarely go deeper than two or three levels.

The EC establishes a classification of digital competences in 5 categories (always in the Internet domain):

• Information processing
• Content creation
• Communication
• Problem resolution
• Security

In each of these categories, the Commission proposes a framework whit three levels of user competence (basic, independent, expert). For each level of competence, standard statements are proposed to help users to perform a self-assessment to establish their digital competence level.

Now, the question that arises in various forums is how to take a step beyond the simple self-assessment of competencies. Questions as “how to perform a search for terms related to digital skills?” are not trivial to answer.

When the skills classification has one or two levels of depth, it is enough to perform a "literal search" of the term to be searched. But what happens if the term tree related to digital skills has thousands of terms hierarchically organized?

For example, imagine that a company needs a very specific profile for a new position in an R&D project. The required profile is a professional with advanced knowledge in MLib Apache Spark library and more than 2 years of experience in streaming Big Data. In addition to these skills - called hard skills - the professional needs to have a series of social skills or soft skills such as public communication ability and teamwork.

How can we find such a profile in a database of 400,000 employees around the world?

A possible solution to these and other issues may be the creation of a digital skills ontology in Europe.

An ontology provides a hierarchical organization of terms (taxonomy) and a set of relationships between them, which facilitates the search - both literal and inferred - of complex terms and expressions. This is already very useful by itself, but if you also combine the formal structure of an ontology with its technical implementation, using a technological tool, you get a powerful technological product. A technical implementation of this ontology would allow, among others, perform the following complex search in an efficient and unambiguous way:

Find a person WITH MLib technical skills that also HAS more than 2 years of experience in streaming Big Data and also HAS the soft skills of teamwork and communication skills at the intermediate level.

With an underlying ontology, in the previous example, all the underlined terms would have a unique identifier, as well as their relations (uppercase). The semantic search engine would be able to identify the previous query, extract the key terms, understand the relationships (WITH, HAS, more, etc.) and execute a search against the employees database, extracting those results that fit with the search.

A good example of using an ontology to perform complex searches in immense databases is SNOMED-CT. It is a standard vocabulary to search clinical terms in patient databases. The clinical domain is especially indicated for the development of ontologies due to the complex structure inherent to the clinical terms and their relationships.

Although there are classic tools and methods of organizing information based on traditional databases and relational models, ontologies and their technological implementations offer higher flexibility, scalability and level of personalization to different subfields.

Precisely the characteristics of flexibility and high scalability become fundamental as open data repositories become increasingly bigger and diverse. The European Open Data portal contains more than 12,000 datasets classified by topics. For its part, the website specializing in data science, Kaggle, hosts 9000 datasets and, annually, they organize competitions to reward those professionals who best analyse and extract useful information from these data. In short, the volume of data available to society is increasing year after year and ontologies are importance as a powerful tool for managing information hidden under that blanket of raw data.

Content prepared by Alejandro Alija, expert in Digital Transformation and innovation.

Contents and points of view expressed in this publication are the exclusive responsibility of its author.