Facebook AI Research (FAIR) has recently conducted a large data study to determine whether social media ties play a role in people's decision to evacuate the affected areas during a hurricane. And, according to the survey, those with the most "social capital" were more likely to take official warnings and leave.
Who knew social media could motivate intelligence in a society?
The FAIR workshop undertook the great data work to determine whether the influence of social media changed a long story about the evacuations. Prior to the emergence of social media, it was found that people who have strong roots in the community are less may be evacuated.
The idea is that these people are obliged to stay because they feel secure in the knowledge that their neighbors, friends and colleagues will help if they are hit by a disaster.
This, of course, is a religion. No amount can prevent a hurricane from harming your property, but an early evacuation will guarantee you that you will not die in that hurricane.
Now, however, there is evidence that social media really motivate people to evacuate during disasters. In particular, those who built "bridges" outside their local community.
According to the FAIR survey:
We found support for this, in particular by observing that links linked to binding capital such as the number of users' friends (first-degree network size) and their network density (grouping coefficient) are associated with a lower probability of evacuation (formerly more stable from the last). We also notice that the bridging and linking bundles, measured by the size of the second-degree network and the number of policies that followed, are associated with a greater probability of evacuation.
The data that Facebook has produced from users affected by hurricanes Harvey, Irma and Maria, show that the "social capital" mentioned is not only necessary for the size of your friends list. This is how you use it.
Those who follow more than two politicians, for example, are among those most likely to evacuate. Facebook states that this can be related to the amount of exposure from "official" sources to information and evacuation directions.
The other important contributing factor was the number of "second-degree" network friends that a user had. These are people outside of your family and local friends. People with stronger secondary networks were more likely to evacuate. This seems to indicate that the more exposure a person has to people outside their local environment during a disaster, the more likely they are to feel comfortable leaving the home.
This is probably true even when there is no disaster.
Some of these sound like common sense, but there are astonishingly little empirical information about evacuation habits. Once people are forced to evacuate, they are scattered. Sometimes they do not return for months or years, and often they are not at all. This makes data collection, under traditional circumstances, hit-or-lose.
But Facebook has access to real-time billions of people in real-time and the ability to measure the size and breadth of their associations all over the world. Cities may lose their citizens (and their data) once they leave the city, but Facebook's international jurisdiction is global. Because of these new elements, we are in unfamiliar territory.
There is certainly room for improvement though. These key results are just the beginning. Soon, Facebook will invite these incredibly broad batters into details:
For example, the analyzes we present will be enriched by adding data about the content associated with evacuating a user consuming social media, including ads, messages from friends, and public officials, to study the impact of specific messages on evacuation behavior that could help in immediate evacuation efforts in the future.
Everything that makes people choose security to "get out" during a hurricane is good. Now, if we could only find a big trick to make it easier for local governments to provide evacuation services for the poor, the disabled and the elderly, we would be facing the real problem.
Tencent to restrict certain mobile players in one hour of play per day