A new study in the journal Science Advances has proposed a low cost method of mapping and predicting storm damage using Twitter. The researchers from Australia focussed particularly on data collected in 2012 around Hurricane Sandy.
The team mapped 10 million tweets from 2 million user and found that there was a greater intensity of activity in areas that came in contact the storm. In order to determine if the increased activity had any correlation with increased damage the team compared their findings with that of FEMA and state governments
The study found “per-capita Twitter activity strongly correlates with the per-capita economic damage“. Their models were actually slightly more accurate than those of FEMA, with the added advantage of being low-cost.
The team seems to have have accounted for any potential problems which data of this nature may create, bots and individual variations for example. They were able to effectively demonstrate the usefulness of the Twitter network in mapping the perceived threats and actual damaging results of the hurricane. Additionally the team suggests that the technique can be used for many different disasters and that many other networks can be used for this purpose.