The humanitarian crisis resulting from the Russian invasion of Ukraine has led to millions of displaced individuals across Europe. Addressing the evolving needs of these refugees is crucial for hosting countries and humanitarian organizations. This study leverages social media analytics to supplement traditional surveys, providing real-time insights into refugee needs by analyzing over two million messages from Telegram, a vital platform for Ukrainian refugees in Germany. We employ Natural Language Processing techniques, including language identification, sentiment analysis, and topic modeling, to identify well-defined topic clusters such as housing, financial and legal assistance, language courses, job market access, and medical needs. Our findings also reveal changes in topic occurrence and nature over time. To support practitioners, we introduce an interactive web-based dashboard for continuous analysis of refugee needs.

Paper Number



Track 15: Student Track