Digital Earth from Vision to Practice: Making Sense of Citizen Generated Content

The vision of Digital Earth put recently forward under the auspices of the International Society for Digital Earth extends the paradigm of spatial data infrastructures by advocating an interactive and dynamic framework based on near-to-real time information from sensors and citizens. This paper contributes to developing that vision and reports the results of a two-year research project exploring the extent to which it is possible to extract information useful for policy and science from the large volumes of messages and photos being posted daily through social networks. Given the noted concerns about the quality of such data in relation to that provided by authoritative sources, the research has developed a semi-automatic workflow to assess the fitness for purpose of data extracted from Twitter and Flickr, and compared them to that coming from official sources, using forest fires as a case study. The findings indicate that we were able to detect accurately 6 of 8 major fires in France in the summer of 2011, with another 4 detected by the social networks but not reported by our official source, the European Forest Fire Information Service. These findings and the lessons learned in handling the very large volumes of unstructured data in multiple languages discussed in this study provide useful insights in the value of social network data for policy and science, and contribute to advancing the vision of Digital Earth.

CRAGLIA Massimo;  OSTERMANN Frank;  SPINSANTI Laura; 

2012-10-16

TAYLOR & FRANCIS LTD

JRC72914

1753-8947,   

https://publications.jrc.ec.europa.eu/repository/handle/JRC72914,   

10.1080/17538947.2012.712273,