Abstract

On May 8, Data Friendly Space will host the session "Integrating NLP into Risk Analysis Methodologies to Drive Robust, Fast, and Resource-Efficient Anticipatory Action" within the context of HNPW 2024.

Anticipatory action is vital in disaster risk management, enabling proactive measures based on predictive analysis to mitigate the impacts of forecasted hazards, thereby protecting vulnerable populations, enhancing preparedness, and building resilience against future crises. Comprehensive, rigorous and regular risk analysis is integral to ensure an effective anticipatory action. It forms the backbone of forecasting and early warning systems by identifying potential hazards and their likely impacts.

This process, however, faces several significant challenges. The rigorous methodology required in risk analysis demands the processing and complex classification of vast amounts of diverse, constantly evolving data, making it a cumbersome and resource-intensive process. This complexity, coupled with the absence of a unified tool that ensures access to all the relevant and structured data necessary to inform risk analysis, often results in the exclusion of local organizations with limited capacity and data literacy from effective anticipatory action, despite their essential role on the ground.

To navigate these challenges, Data Friendly Space proposes the integration of contrasted and rigorous risk analysis methodologies in a set of advanced tools that would consolidate and pre-process all necessary qualitative and quantitative data to carry out risk analysis into a single interactive and open space, to later assist the analysts in their process of assessing risks. The key factor to accomplish this is DFS’ expertise in Artificial Intelligence (AI) and Natural Language Processing (NLP) applied to the humanitarian and development sectors. DFS builds generative AI tools to access and extract complex and high varieties of quantitative and qualitative data, clean, structure them and retrieve required information to support fast and rigorous analysis at emergency response and anticipation such as impact and response information retrieval and monitoring, humanitarian virtual assistance for efficient and near real-time search and query.

The session is designed for a wide array of attendees, encompassing humanitarian and development organizations involved in different types of analysis, individuals interested in the use of AI and NLP for data structuring and analysis, professionals engaged in risk analysis and anticipatory action, as well as donors and stakeholders seeking innovative data management and analysis solutions. This diversity ensures a rich exchange of ideas and collaborative learning opportunities in the fields of disaster risk management and technology application.