webpage portfolio for open source GIS work
As defined by Crawford & Finn (2014), a disaster is a disruption of regularity in a particular space and time, though is far more dependent on existing structures of power social patterns. As illustrated by the current adage about COVID that “we are all in the same storm, but not on the same boat”, it is recognized by critical scholars that prior social conditions will impact our experience, perception, and outcome in response to a disaster. Crawford & Finn (2014) explore the ontological, epistemological, and ethical concerns with volunteers geographic information (VGI), particularly concerning sources of uncertainty presented by these concerns. With ontological issues in VGI, people experience disasters differently, often based on personal experience, imagined experience, lived experience, and ongoing traumas in their lives.
In terms of epistemological concerns around how data is created and collected, there is a great deal of uncertainty and inconsistency of who posts, who has access to internet and posting capabilities, and how posts are promoted through internal algorithms of posting platforms. For example, on Twitter, there is a great deal of “performative posting,” along with numerous bots and institutions who use the site. Amongst individual users, there is a particular skew towards younger users living in urban settings. For researchers looking to use Twitter data to understand disasters, it provides an incomplete picture of how disasters affect different communities, as there are a lot of people (older, more rural, less wealthy) who are far less likely to use Twitter than their urban counterparts. These issues of uncertainty in representation expand into how data collection platforms use internal algorithms to influence the retweeting and sharing of particular posts, not providing an organic look into networks of connectivity amongst users. On top of these data concerns, researchers often will filter data based on the language in which they are working or will only look at disasters in certain timeframes, not in the long-term social contexts.
Lastly, there are numerous ethical issues surrounding privacy and if people posting during a disaster have consented to having their vulnerabilities used in research. Especially in instances with VGI submitted by people suffering from a disaster, there must be some two-way directionality of data submission, where those providing data are directly compensated or supported, not just told that their work is going towards “higher-level” solutions that will mitigate future disasters. With data scraped from social media platforms, and not explicitly collected from open-source programs where people are knowingly providing data about a disaster, users experiencing a disaster may have evolving feelings towards the privacy of their data during and after the disaster. While users may not place a huge value on data privacy when trying to locate family members in the midst of a disaster, they may have differing opinions when their posting, when at their most vulnerable, is taken and aggregated for scientific research (Crawford & Finn 2014). In my ideal world, governments around the world would create legislation barring the collection of personal information by social media platforms (and sharing of this data) without the expressed consent of users. While arguments about the “greater good” over informed consent may be used by some, I firmly believe that there is a responsibility on the part of researchers to use existing data with caution and to be transparent about the ethical complications of using big data without the consent of participants. To not consider the consent and ethical complications of research would go against the strict guidelines of participant consent that Institutional Review Boards purport to uphold.