Name: Ramesh De Silva
Title: Web GIS Application Integrated with Social Network to Disseminate Disaster Information Based on User Preferences and Location Information
Due to the vast development in information andcommunications technologies, the availability of information on Internet has grown rapidly and users are overwhelmed with information. Disaster alert web portals address this problem by allowing people to define their interested regions, as they may need not to know about disasters happening all around the world.
Though disaster notification systems allow users to define their area of interest, it has its own limitations. Assume that theuser travels away from the area of interest. The system would not track the new geographical location of the user. If a disaster occurs at a close vicinity tothe new location of the user, the system would not send any alert.
To address this need, the geographical movementsof users need to be tracked by some means. Although it is possible todynamically track the geographical location of users through mobile cellular devices, mobile service providers do not disclose such information without bounds.
On top of provision to define area of interest,the proposed system uses a social network integrated approach to dynamically get the location information of the users. Also, location information of the concerned family members or friends of the users can be acquired by this mechanism. Users get access to this application via Facebook Open ID login and define their area of interest and concerned friends from the facebook friendslist. The system periodically retrieves location information of the users and their concerned friends via Facebook Graph API.
This application is designed to send earthquake alerts. New earthquake events are fetched from USGS earthquake notification system. When a new earthquake event is observed, the system checks the affected area of the earthquake against area of interest of each user and the location information of the users and their friends. Notifications are sent only topossibly affected users.
Performance remains as a key concern in disaster alert systems. Various performance tests are performed to check the behavior of the system. Spatial indexes and spatial clustering techniques are used in the tests.
The proposed system works as a fully automated system to fetch earthquake feeds and notifying the affected users. The affected users are efficiently found by analyzing the area of interest marked by theuser and the location information obtained from Facebook Graph API with the earthquake hit area.