Digital Communities and Natural Disasters

Key Words: Natural Disaster, Data Mining, VGI, Big Data

While their immediate impact in on physical communities, natural disasters are connected to the digital world as well. This connection has grown as the institutions which respond to disasters, such as first responders and governments, have become more aware of social media's influence.

Awareness
Many relief organizations and governments now use social media regularly to spread awareness about emergent disasters. This is done in addition to sharing that information through radio, television, and other online spaces. Social media allows citizens to engage with information, while the creation and use of official hashtags makes categorizing and searching for information easier.

Data Mining Social Media
The wealth of information provided by the public on social media can be used to learn about natural disasters. An algorithm is trained with machine learning to recognize posts that relate to a disaster, and potentially to interpret the meaning and evaluate the quality of information as well. That might mean identifying flooded roads in pictures or finding Tweets related to an earthquake. Information the algorithm finds can be used to identity unknown events, sometimes much earlier than other methods could identify them, or to learn about ongoing or recent events. To be useful, these algorithms must be able to determine where an event is happening. Some posts might have a marker that shows the exact position or provide the name of the location, but posts will often only vaguely reference the area. They might state the name of the city it is in or near, or say nothing at all about location. To attempt to narrow down the range, algorithms may scan a user's history and their site friends' histories.

This topic has received significant academic discussion, but its adoption has been slower. The extent is currently hard to determine, but it is definitely used by some organizations. The most popular social network to scan for disaster information is Twitter; it can provide accurate time and location data and since it is mostly text, it is easier to interpret.

If you find yourself in a disaster situation, it may be worth considering how a response organization could use your social media posts. To aid in their response, you could use specific locations and use relevant hashtags.

Data mining raises concerns regarding privacy and Big Data. The people who provide the information that is mined will likely never know it was taken, and are not asked for consent or compensated. While this use of data mining has only positive effects, the question remains if data mining needs to be used for negative effects to be immoral.

Volunteered Geographic Information (VGI)
Another tool for addressing natural disasters that takes advantage of digital communities is the use of Volunteered Geographic Information, or VGI. This is any location information a person voluntarily shares online, such as in a citizen mapping project like Wikimapia. Unlike in data mining, VGI is shared with end users intentionally. There are several cases where maps and information built by online communities have helped in disaster response. In 2010, a devastating earthquake struck Haiti. Haitians and many groups outside the country, especially in the United States, began mapping the effects of the quake and the area around it. Using publicly available satellite imagery, they made the best street map of Haiti that existed at the time. This information allowed response teams to better locate people in danger and saved lives. The Haiti effort relied on Ushahidi, an open-source app designed for collecting location data. Other success stories include the 2008 Tea Fire in Santa Barbara, California and tracking Lyme Disease in the Netherlands. When a disaster changes the landscape, or if there are not detailed maps of a place to begin with, citizens can take the initiative in getting resources and first responders where they need to go.