Bridging the Urban Divide: Adaptive Cross-City Learning for Disaster Sentiment Understanding

TL;DR

This paper presents an adaptive cross-city learning framework that improves disaster sentiment analysis from social media by incorporating mobility data and city similarities, leading to more accurate and equitable insights during natural disasters.

Contribution

It introduces a novel multi-modal, city-aware learning approach that enhances disaster sentiment understanding across diverse urban contexts, addressing biases in existing models.

Findings

Achieves state-of-the-art performance on wildfire sentiment analysis

Reveals geographically diverse sentiment patterns during disasters

Shows that combining behavioral and textual data improves accuracy and fairness

Abstract

Social media platforms provide a real-time lens into public sentiment during natural disasters; however, models built solely on textual data often reinforce urban-centric biases and overlook underrepresented communities. This paper introduces an adaptive cross-city learning framework that enhances disaster sentiment understanding by integrating mobility-informed behavioral signals and city similarity-based data augmentation. Focusing on the January 2025 Southern California wildfires, our model achieves state-of-the-art performance and reveals geographically diverse sentiment patterns, particularly in areas experiencing overlapping fire exposure or delayed emergency responses. We further identify positive correlations between emotional expressions and real-world mobility shifts, underscoring the value of combining behavioral and textual features. Through extensive experiments, we demonstrate that multimodal fusion and city-aware training significantly improve both accuracy and fairness. Collectively, these findings highlight the importance of context-sensitive sentiment modeling and provide actionable insights toward developing more inclusive and equitable disaster response systems.