Volunteered Geographic Information and Uncertainty
One of the most prevalent sources of uncertainty in volunteered geographic information (VGI) is the demographic skew of the contributors. Depending on the dataset, the people involved may skew towards the more educated, the more urban, the young, or the wealthy. As Crawford and Finn (2014) write with respect to tweets about Hurricane Sandy, “certain groups are well represented and others are absent” (499). Technology ownership and use generally skews towards the younger, wealthier, and more urban. If the VGI used in research is not representative of the population, it can distort perceptions of disasters and humanitarian crises.
The potential (dis)honesty, (un)reality, and (in)accuracy of VGI is also an important consideration when it comes to uncertainty in research on disasters. While a researcher might have little reason to believe people are deliberately misrepresenting disasters or other phenomena, people do lie on the internet. They also exaggerate; Crawford and Finn note that “some news organizations and journalists use disaster events to increase their follower count and their readership by amplifying the most dramatic images and updates” (496). Furthermore, accidental inaccuracy is rampant.
The primary ethical concerns when working with VGI ought to be the privacy and consent of those who volunteer the information. Oftentimes, public messages like tweets are not intended to be disseminated widely, and might contain sensitive information. Not to mention that Twitter users are not given a real choice to opt out of the use of their data, so that they have not truly consented to its use in research. As to the ethics of uncertain data, there is an obligation on the part of the analyst to be transparent and honest.
… … …
References
Crawford, K., and M. Finn. 2014. The limits of crisis data: analytical and ethical challenges of using social and mobile data to understand disasters. GeoJournal 80 (4):491–502. DOI:10.1007/s10708-014-9597-z