Goodchild and Li (2012) outline crowd-sourcing, social and geographic approaches to quality assurance for volunteered geographic information (VGI). Representing an increasingly important resource for data acquisition, there is a need to create and interrogate the frameworks used to accept, query or reject instances of VGI on the basis of its accuracy, consistency and completeness.
The authors argue that VGI presents a distinct set of challenges and considerations from other types of volunteered information. For example, Linus’s Law—that in software development, “given enough eyeballs, all bugs are shallow”—may not apply as readily to geographic facts as it does to other types of information. Evaluators’ “eyes” scan highly geographic content selectively, with exposure of geographic facts varying from the very prominent to the very obscure.
To me, it is unclear why this disparity is unique to geographic information. The direct comparison between Wikimapia and Wikipedia may be inappropriate for contrasting geographic/ non-geographic volunteered information, since their user/ contributor bases differ so markedly. I might actually advance in the opposite case; that the fact that geographic information is all connected by locations on the surface of the earth makes it more ‘visible’ than, for instance, an obscure wikipedia page on an isolated topic.
The authors call upon further research to be directed towards formalising and expanding geographic approaches to quality assurance. These approaches seek to verify VGI using external information about location and by applying geographic ‘laws’. In my opinion, this provides an interesting strategy that is relatively unique to geographic information. Through geolocation, any instance of VGI could be linked to other geospatial databases, and could potentially be accepted or flagged on the basis of their relationships to other nearby features or variables. Elements of this process could be automated through formalisation. This approach will of course come with its own set of challenges, such as potential feedbacks generated by multiple incorrect sources reaffirming inaccurate information.
-slumley