Browse Prior Art Database

Early dissaster detection using social network actitivty frequency analysis

IP.com Disclosure Number: IPCOM000245151D
Publication Date: 2016-Feb-14
Document File: 4 page(s) / 43K

Publishing Venue

The IP.com Prior Art Database

Abstract

When a natural disaster occurs time is of the essence. The sooner the disaster is detected the sooner it can be properly managed and mitigated. We propose a simple and prompt method for finding such regional disasters based on sharp changes in the frequencies of messages sent.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 4

Early dissaster detection using social network actitivty frequency analysis

This invention addresses the problem of early disaster detection using social media .

    When a naturaldisaster occurs time is of the essence. The sooner the disaster is detected the sooner it can be properly managed and mitigated. Some disasters, such as storms and earthquakes, can be predicted to a certain extent. However, not all disasters can be predicted all the time. For example, if the disaster is of relatively small scale, occurs in a country without a complicated disaster warning system, or in cases of massive accidents and terror attacks that are by nature unpredictable.

    One of the side effects of a massive disaster is that immediately after its occurrence social-media activity in the disaster's area drastically declines. Infrastructure can be damaged; people are busy recovering from the disaster, locating their loved ones, or escaping unsafe locations. They will only regain activity in the social media once they are safe.

    This invention proposes a geospatial apparatus for early disaster detection by using social media volume and content analysis. It monitors social-media activity for sharp declines and then analysis the content in those areas to detect the nature of the disaster and prevent false alarms .

This method operates in three stages as follow:

    First stage: Detect possible disaster stricken areas.The area of interest is divided to a grid structure. Each slot corresponds to a sub-area of the original area and its activity volume is monitored. When activity frequency in a location drops below a predefine threshold (absolute or relative, depends on the specific implementation) a disaster is assumed to have occurred. Second Stage: Group all suspected grid slots to a cluster.

This is a relatively easy programming task.

Third S...