Data are a key driver in the modern era. Combining data from different domains brings new insights and adds further dimensions to scientific research, but data on their own rarely explain themselves. Without context, data are not meaningful and can lead to erroneous conclusions. Making sense of data is at the heart of recent research led by Dr Nicholas Nicholson at the European Commission’s Joint Research Centre and Dr Iztok Štotl at the University of Ljubljana, to build a standardised data-contextualisation framework called SOLICIT. Read More
A common problem is encountering a dataset with little or no explanation attached. Descriptors of data variables – essentially, the metadata – are often minimal and ambiguous. Users are then left guessing – and important nuance is easily lost. The same is true when artificial intelligence learning models are used to combine data.
This is why a framework such as SOLICIT is key. SOLICIT’s contextualisation components describe what data mean, where they come from, and how they can be used – helping both people and automatic learning models understand them correctly.
To make this approach work across many fields, Nicholson and Štotl’s research shows the need to integrate standard metadata models with foundational ontologies. An ontology is a shared model of concepts and their relationships, allowing contextualisation constructs to be understood consistently across domains. Foundational ontologies provide common building blocks that can be reused instead of reinvented.
Nicholson and Štotl’s work builds on established foundations such as the Basic Formal Ontology and the Common Core Ontologies. SOLICIT integrates these ontologies with the ISO/IEC 11179 metadata registry standard, embedding standardised metadata concepts directly into the ontology layers. The standard’s metadata model allows for sharing of metadata elements by breaking them into three constituent parts: the Object Class (what kind of thing is being described), the Property (what aspect of it is measured), and the Value Domain (the allowed values and measurement units).
The contextualisation components implemented within SOLICIT provide the information necessary for reliable interpretation and processing of the data described via a stable actionable unit with a persistent identifier, enabling full compliance with the FAIR (findable, accessible, interoperable, and reusable) data principles.
This example shows part of a contextualised indicator measuring age-standardised cancer rates. Rates are normalised to the European standard population.
Rates are given as ratios. The numerator and denominator are described in terms of elements of the ISO metadata registry standard. Further qualifications can be added. Dashed lines point to entity values. Further context can be added too, empowering users to understand what is behind bare numbers. The context for data quality, or indeed for any context, can be elaborated to any degree necessary.
Although SOLICIT was originally designed to contextualise indicators, it turns out to be far more flexible – able to contextualise any type of dataset and integrate data quality frameworks in a consistent manner. SOLICIT tracks the source of data throughout the processing chain and any final overall composite data quality score can then be further inspected to understand how the score was attributed.
As Nicholson and Štotl’s research makes clear, combining structured metadata with ontologies offers a practical path toward clearer, more trustworthy, and more reusable data for everyone.
