[Paper Review] Earthquake Impact Analysis Based on Text Mining and Social Media Analytics

Zhe Zheng, Hong-Zheng Shi+3 more|arXiv (Cornell University)|Dec 12, 2022

Public Relations and Crisis Communication4 citations

TL;DR

This paper proposes a text mining and social media analytics framework to enable early earthquake impact assessment using Sina Weibo microblogs. By analyzing trending keywords, microblog volume, sentiment shifts, and rule-based classification, the approach successfully differentiates disaster severity—demonstrating that public sentiment trends and volume spikes can predict social impact earlier than traditional methods, aiding emergency decision-making.

ABSTRACT

Earthquakes have a deep impact on wide areas, and emergency rescue operations may benefit from social media information about the scope and extent of the disaster. Therefore, this work presents a text miningbased approach to collect and analyze social media data for early earthquake impact analysis. First, disasterrelated microblogs are collected from the Sina microblog based on crawler technology. Then, after data cleaning a series of analyses are conducted including (1) the hot words analysis, (2) the trend of the number of microblogs, (3) the trend of public opinion sentiment, and (4) a keyword and rule-based text classification for earthquake impact analysis. Finally, two recent earthquakes with the same magnitude and focal depth in China are analyzed to compare their impacts. The results show that the public opinion trend analysis and the trend of public opinion sentiment can estimate the earthquake's social impact at an early stage, which will be helpful to decision-making and rescue management.

Motivation & Objective

• To develop a real-time method for assessing earthquake impact using social media data.
• To identify early indicators of social disruption through microblog volume and sentiment trends.
• To compare the social impact of two earthquakes with identical magnitude and depth using text-based analysis.
• To support emergency management by providing timely, data-driven insights from public discourse.

Proposed method

• A web crawler collects disaster-related microblogs from Sina Weibo based on predefined keywords and event triggers.
• Data preprocessing includes removing noise, filtering non-relevant content, and normalizing text for analysis.
• Hot word analysis identifies frequently mentioned terms related to damage, casualties, and rescue efforts.
• Time-series analysis tracks the volume of microblogs and public sentiment (positive/negative/neutral) over time.
• A keyword- and rule-based text classification system categorizes microblogs into impact types such as infrastructure damage or human casualties.
• Comparative analysis is performed on two recent Chinese earthquakes with identical magnitude and focal depth to assess relative social impact.

Experimental results

Research questions

• RQ1 Can social media activity trends serve as early indicators of earthquake impact severity?
• RQ2 How do sentiment shifts in public microblogs correlate with actual disaster consequences?
• RQ3 Can text mining distinguish between earthquakes of similar seismic parameters but different social impacts?
• RQ4 To what extent can public sentiment and microblog volume predict the scope of disaster response needs?

Key findings

• Public opinion sentiment trends and microblog volume spikes provided early signals of earthquake impact, detectable within hours of the event.
• The method successfully differentiated the social impact of two earthquakes with identical magnitude and depth, revealing higher public concern and longer-lasting discourse for the more severely affected region.
• Hot word analysis identified key impact indicators such as 'collapsed buildings', 'rescue', and 'power outage' in real time.
• The combination of sentiment tracking and volume trends proved more effective than magnitude alone in estimating social disruption.