Sharing photos online is a vital resource

When a natural disaster hits, people are quick to pull out their smart phone and take a picture to share with others. It is this behaviour that professors from Warwick Business School now say is a valuable resource for the government and commercial companies.

Tobias Preis, associate professor of behavioural science and finance, and Suzy Moat, assistant professor of behavioural science at the UK business school, compared pictures of Hurricane Sandy uploaded on to Flickr, the photo-sharing social media site, with data from the US weather service. Focusing on October and November 2012 only, they found a strong link between the number of photos uploaded and atmospheric pressure in New Jersey.

The research, titled Quantifying the Digital Traces of Hurricane Sandy on Flickr and published in Scientific Reports, shows, for example, that the highest number of pictures posted were taken in the same hour in which Hurricane Sandy hit New Jersey.

“Plotting the data revealed that the number of photos taken increased continuously while ‘Sandy’ was moving toward the coast of the US. This study would suggest that in cases where no external sensors are available, it may be possible to use the number of Flickr photos relating to a topic to gauge the current level of this category of problems,” says Prof Moat.

“Flickr can be considered as a system of large scale real-time sensors, documenting collective human attention.”

While both Prof Moat and Prof Preis agree this type of big data is not always entirely reliable, they do strongly believe it is an improvement to previous data sources such as surveys, and could help policy makers measure the impact of disasters.

This type of technology offers scope and opens people’s minds to linking the online activity of humans to real world activity, explains Prof Moat. In addition, the two professors believe social media data could be a useful commercial tool. Insurance companies, for example, could use the photos to help them assess the damage of individual properties, which is normally a lengthy process, says Prof Preis.