Social Sensing and Social Media for Crisis Management

Social media, citizen-generated digital data, and the broader crisis infosphere are important sources of information during disasters, humanitarian emergencies, conflicts, public-health events, and slow-onset and climate-related crises. Building on ISCRAM’s long-running Social Media for Crisis Management track, this track broadens the scope to social sensing and infosphere intelligence: collecting, interpreting, integrating, and responsibly using observations generated by people and communities through social media, crowdsourcing, mobile and participatory platforms, alongside information from news services, sensors, satellite imagery, emergency communication platforms, and other digital sources.

The track welcomes technical, socio-technical, organizational, and human-centered research on social sensing and infosphere-based crisis intelligence for situational awareness, crisis communication, decision-making, community resilience, and recovery. In line with the ISCRAM 2027 theme, “Supporting First Nations and Remote Communities,” we particularly encourage work addressing culturally appropriate communication, community priorities, connectivity, geographic isolation, language, unequal representation, trust, and access to timely emergency information.

Topics of Interest

Topics include, but are not limited to:

• Social sensing in crisis contexts: How social media, crowdsourcing, and other citizen-generated data are produced, interpreted, and used across disasters, conflicts, climate-related emergencies, public-health events, and slow-onset crises.
• AI and large language models for crisis social sensing: Machine learning, foundation models, and LLMs for information extraction, summarization, classification, translation, question answering, decision support, and crisis mapping, including reliability, hallucination, bias, explainability, and responsible use.
• Multimodal crisis information analysis: Extracting actionable information from text, images, video, audio, and other modalities for event detection, geolocation, damage assessment, and information fusion.
• Information quality and information overload: Credibility assessment, rumor and misinformation detection, prioritization of high-value information, uncertainty management, and trustworthy crisis communication.
• Crisis communication and societal sensemaking: Public narratives and sentiment during crises to understand public responses and inform crisis communication.
• Social sensing for decision-making: Incorporating social sensing into crisis decisions, identifying emerging concerns and changing information needs, and tracking evolving public responses.
• Hazard impact extraction and quantification: Infosphere-based assessment of physical damage, economic losses, and infrastructure disruption; near-real-time severity classification; automated loss estimation; cascading effects; and integration with physical loss models.
• Vulnerability and exposure mining: Using diverse data and AI/ML methods to assess population vulnerability, social exposure, and systemic risk, including socially vulnerable groups, critical infrastructure dependencies, transient populations, and changing exposure patterns.
• Multi-source information fusion: Integrating social sensing with emergency alerts, news, weather and hazard data, satellite or sensor data, call-center records, maps, volunteered geographic information, and other sources.
• Human-centered design and operational adoption: How emergency managers, humanitarian practitioners, community groups, and affected populations use social sensing systems, including information needs, workflows, barriers, trust, usability, and participatory or co-design approaches.
• Case studies, systems, datasets, benchmarks, and best practices: Empirical studies, operational systems, datasets, benchmarks, reproducibility studies, and lessons from developing, deploying, or assessing social sensing technologies for crisis response and management.

← Back to Call for Papers and all tracks