Project crAIRsis

Characterizing crises-caused air pollution alterations using an artificial intelligence-based framework

ABOUT crAIRsis

Acronym crAIRsis
Duration  January, 3rd 2024 – January 2nd 2027
Grant No.  7373
Funded under the Science Fund of the Republic of Serbia, Program PRISMA
Project budget  € 284,907.86
Coordinated by Prof. Milan Tuba, Singidunum University, Belgrade, Serbia
Participants Singidunum University, Belgrade, Serbia, and the Institute of Physics Belgrade

Read More

MAJOR CHALENGES

Integration of Diverse Crisis-Related Data: The crAIRsis project integrates a broad spectrum of data sources, including governmental, organizational, and public datasets. This effort focuses on optimizing the utility of public data and tackling complex data management challenges. By harmonizing data with varying characteristics, the project aims to create a structured dataset that supports an AI-driven analytical framework, enhancing our ability to model air pollution changes during crises and improving predictive capabilities for decision-making.

Advanced Geospatially-Fused Big Data Analysis: Employing sophisticated geospatially-fused big data analytics, the crAIRsis project seeks to uncover patterns of pollutant behavior and assess the complex impacts of crises on air quality. This approach combines high-resolution spatial data with extensive temporal datasets, providing a comprehensive view of how air pollution dynamics respond to various crisis conditions. This analysis is crucial for identifying the factors that influence air quality and developing effective air pollution management and crisis response strategies.

Modeling the Crisis Life Cycle: The project develops models to capture the entire lifecycle of a crisis—from inception through recovery. Using advanced machine learning techniques and AI frameworks, these models are designed to predict early signs of crisis impact, monitor progression, and forecast long-term effects on air quality. Adaptable to different stages of a crisis, these tools are vital for stakeholders managing air quality under evolving conditions.

Pioneering Environmental Science through the crAIRsis Project: The crAIRsis project introduces a groundbreaking approach to environmental science by focusing on air pollution in crisis situations. Leveraging state-of-the-art modeling techniques and extensive big data analysis, the project aims to transform insights into the effects of crises on air quality. This innovative framework significantly enhances decision-making, crisis management, and scientific research with in-depth, data-driven insights.

KEY ASPECTS OF THE crAIRsis CONCEPT

Global Data Collection: Utilizing an extensive dataset covering air quality across Europe, the project enables detailed analysis of region-specific conditions during crises.

Contextual Data Analysis: Integrating diverse datasets helps to contextualize air pollution within specific crisis scenarios, providing a detailed characterization of air pollution phenomena.


Holistic Modeling Approach:By examining the interconnected impacts of crises on air quality, the project identifies crucial environmental factors that influence air pollutants.


Synergistic Transdisciplinary Methodology: Merging data science with environmental science, crAIRsis enhances understanding of how air pollution evolves during crises, promoting a comprehensive, transdisciplinary research approach.

PROOF OF CONCEPT: Analyzing COVID-19’s Impact on Toluene Concentrations

Focusing on the effects of COVID-19 on toluene levels in Europe, the project's proof of concept incorporates:

1. Comprehensive Data Collection: Gathering data on toluene concentrations alongside meteorological, mobility, and policy data during the COVID-19 crisis.
2. Advanced Modeling Techniques: Using Machine Learning and SHAP analysis to understand toluene's behavior under crisis-induced environmental conditions.

These initial findings have unveiled unique patterns of toluene behavior, deepening our understanding of pollutants under varying conditions. As crAIRsis advances, it continues to refine its methodologies in data processing and ML application, ensuring robust and comprehensive outcomes that reinforce its pivotal role in environmental science during crises.