Towards a comprehensive ontology for disaster risk management
Users need tools to help them filter and analyze increasingly large volumes of information: and often they don’t have strong IT or data science skills. Still, they need tools to make advanced searches, get visual analytics and produce reports. And when it comes to policymakers and regular citizens, they expect highly interactive and engaging tools that provide good user experience.
Designing these tools and the underlying exploratory user experiences requires a deep understanding of the stakeholders — their specific language and jargon— and the workflows needed to process and rearrange information to create new data. Finding raw knowledge, which is often scattered across different organisations, each with their own “system of systems”, can be challenging. But speaking from experience, it is certainly a challenge within grasp.
Overall, there is a need to link data and information across various information systems and tools through a collaborative process, increasingly involving different types of expertise. That’s why knowledge management tools should serve users with different backgrounds, interests and mandates, and help them work together.
Disaster Risk Management Taxonomy
In disaster risk management, information often comes under the shape of knowledge graphs with different detail levels across multiple disciplines. And not only humans should be able to analyse and digest this information: it also needs to be fit for machine learning so that the prediction power of algorithms keeps improving.
This is the effort on which DRMKC embarked at the end of 2021, together with UNDRR. The interested Directorates-General of the European Commission started a working group to co-organise the existing knowledge base in a comprehensive taxonomy. Such taxonomy aims to cover not only the components of risk (hazard, exposure and vulnerability/resilience), but also DRM stages and processes (i.e. risk assessment and disaster loss data inventory).
In addition, we aim at integrating terminologies on hybrid threats, megatrends (i.e. long-term global driving forces observable in the present and likely to continue having a significant influence for a few decades) and other relevant terminologies connected to the challenges of today’s society. For example, dedicated terminologies on climate change and conflicts will also be integrated in the taxonomy.
The DRM taxonomy will also include links to EU funded projects, publications, recommendations, datasets and learning materials. It is being adopted as a reference for the ongoing developments of the Union Civil Protection Knowledge Network web platform and will be linked with work on EIOS (Epidemic Intelligence from Open Sources) and the World Health Organization Hub for Epidemic and Pandemic Intelligence.
A prototype version of the DRM taxonomy, based on the conceptual schema shown in the figure, has been presented at the 7th Civil Protection Forum (Brussels, 28-29 June 2022) to approximately 80 participants of 27 countries.
The approach followed in the taxonomy is both top-down (definition of an ontology/conceptual framework) and bottom-up (reviewing/merging existing terminologies, e.g. from EU, UN, WHO). It uses FAIR vocabularies (Findable Accessible Interoperable Reusable and is based on SKOS - Simple Knowledge Organization System- and/or OWL – Web Ontology Language). From this perspective, the approach will allow for machine learning works (e.g. automatic classification of content, pattern discovery). The DRM taxonomy under construction is based on VocBench (EU Publication Office tools, https://op.europa.eu/en/web/eu-vocabularies/online-tools).
Link: https://showvoc.op.europa.eu/#/datasets/JRC_DRMKC_Vocabulary/data
