JANZZ classifiers: The ultimate multilingual tools for easy and automated classification of your occupation data.

JANZZ classifiers analyze complex sets of occupation-related data such as job title, skills, function, or industry, and annotate them with intelligent and standardized meta-data. Over 100 official occupation classification systems, such as O*Net, ISCO-08, ESCO, BO&C, ASOC or SSOC 2015, and over 60 other standardized reference systems are available, as well as up to 40 supported languages and dialects. Through precise classification and standardization, our patented classifiers transform your data into smart data: comparable, validated datapoints, enriched with actionable meta-data. Creating the ideal starting point for further processes such as benchmarking, matching, or statistical analyses, as well as the essential basis of any well-performing solution based on AI and occupational data.

Using semantic technology for precise and intelligent classification

JANZZ classifiers combine the strengths of JANZZ.technology’s key technologies: the knowledge about occupations and skills that is stored in the world’s most extensive ontology / knowledge graph of occupation data (JANZZon!) and the semantic job matching engine that enables an intelligent matching of structured and unstructured data (JANZZsme!). The perfect interplay of these technologies allows for precise and transparent classification of large volumes of data. Importantly, the classifications are not only based on a few keywords in your data set but rather on the analysis of multiple dimensions of your data and their relations, for instance industry, job description and function.

Following the OECD Principles on Artificial Intelligence (AI), JANZZ classifiers are perfect for organizations looking to generate actionable intelligence by transforming large volumes of (unstandardized) occupation-related data into truly smart data:

  • Software providers (HR software, ATS, HCM software)
  • System integrators
  • International corporations
  • Government organizations, in particular public employment services (PES) and national statistical institutes
  • Professional networking sites

Choose the perfect classification tool for your skills and job data

JANZZ offers five different classifiers, depending on your use case.

TITLEclassifier!

This is the ideal choice if you simply need a no-frills solution to classify job titles according to your favorite classification systems, say for statistical purposes. Simply enter a job title in one of the supported languages and the classifier will return the associated classification code and standardized term in each desired classification system or taxonomy.

OJAclassifier!

If you want to extract and classify job titles from job postings, this Online Job Advertisement classifier will get it done. OJAclassifier! detects the job title in a posting if available, or determines the closest match based on the job description if no job title is explicitly mentioned. It then returns a list of most probable job titles and their associated codes in each classification system you specified.

MEJPclassifier!

If you are interested in more granular information such as the associated skills and responsibilities of jobs, our Multiple Entity Job Posting classifier is what you need. MEJPclassifier! can parse a job description to extract and classify any information from categories such as occupation, hard or soft skills, experience, qualifications, industry, contract type and many others into a set of predefined classes.

SECclassifier!

If you are looking for a tool to restructure job postings of improve the results of your parser, use this sectionizing classifier. Unlike the other classifiers, this tool does not classify occupational data as such. Instead, it divides job postings into several sections which it then classifies according to their content, i.e., title, company description, offer, duties, etc.

SKILLSclassifier!

If your focus is on skills and you want to significantly improve the skills matching in your recruiting software or talent marketplace, increase the outreach of your job postings, or generate clean, validated, contextualized data for use in official statistics and policymaking, this is the tool you need. This classifier detects the semantic context of each skill to verify its meaning in each occurrence and translate it into a standardized, unambiguous synonym together with the required level of knowledge and related skills.

All JANZZ classifiers are multilingual and deliver outstanding results tailored to your individual requirements. They are typically accessed via the standard JANZZjobsAPI and can easily handle large data volumes with impressive response times – regardless of whether your data is structured or unstructured. In special cases, we can also offer on-premise setups or dedicated, ultra-performant AWS solutions for very large volumes, if GDPR/CCPA compliance is ensured.

If you too want to generate meaningful and actionable insights for your organization easily and effectively, feel free to contact us via email or via our contact form.

OJAclassifier! – use case: automatic classification of job offers for a PES

Background situation
For years, the employees of an EU state public employment service processed several thousand job offers by hand each month to assign each one the correct code from the national classification system as well as, for statistical purposes, the ISCO-08 code. This largely manual process involving hundreds of thousands of offers each year was not only very expensive, but also very labor- and time-intensive. The process was also extremely susceptible to human error: Many job descriptions were regularly assigned incorrect codes because they were not, or only partially, interpreted correctly and with major discrepancies between the interpretations by individual employees.

Improvements made possible by JANZZ.technology
Thanks to JANZZ’ solutions, the client was able to save public funds while simultaneously considerably reducing processing times. The new procedure is as automated as possible and guarantees largely error-free allocations within the two classification systems.

Solution and methodology

  • Detailed discussions with the project managers allowed for a clear conceptualization of the process regarding the type of data structuring and areas of potential improvement in the context of data entry.
  • The public employment service parses all job offers in advance to provide partially structured text components for further processing. While this pre-parsing is helpful, it is not essential for the processing by JANZZsme!.
  • The data is processed by JANZZsme!, in conjunction with JANZZon!, which accesses background and contextual knowledge on the various job titles to ensure that the correct codes are allocated.
  • JANZZsme! then assigns the codes from the national job classification system as well as  the currently essential international ISCO-08 code to the job offers.
  • If a code cannot be assigned automatically due to unclear or overly general job titles or because of contradictory information in the texts, these cases are flagged and separated by JANZZsme! to be checked manually and allocated by experts.

Accomplishment and results for the client

JANZZ’ solutions reduced the processing times and costs substantially and sustainably for a previously very labor-, time- and cost-intensive process with a high susceptibility to errors. The allocations made by JANZZsme! were on average around 42% more accurate, allowing for improved statistical analysis. Less than 15% of the results had to be flagged for manual checks and subsequent allocation. And since then, our classifiers have been continuously refined and improved. The current flag rate is less than 5%.