In the first post in this series on humanitarian open data and affected populations, I mentioned that there really wasn’t any consensus among humanitarians as to their sector’s conceptualization of the relationship between open data and affected communities. However, it is also important to note that, despite the absence of complete agreement, there do exist some notable themes and commonalities that have surfaced in our conversations with various NGOs and inter-governmental agencies.
Perhaps the themes that ought to be examined first are those that challenge the assumption that open data is a widely accepted and acceptable concept within the humanitarian community. While it is undeniably true that the idea of open data is becoming better recognized and prominently pursued in parts of the humanitarian community, this belief is not universal nor without serious complications.
In our interviews, this reality became most noticeable through the theme of intra-organizational contention. Despite the growing amount of consideration of and enthusiasm for the potential of open humanitarian data mentioned in this topic’s opening post, there still appear to exist considerable rifts within some humanitarian organizations as to the actual importance of both open data and simply ‘big data’ in general. According to one representative of a private humanitarian organization, many organizational systems view data from a mindset that is “stuck in the 2000s,” while a representative from a different organization characterized his efforts at embracing open data as “swimming against the current.” Although certainly not all organizations struggle with this baseline recognition of open data’s general potential, it was stressed by many in the humanitarian community that a divide often exists between those working on data-focused initiatives and their counterparts (and often superiors) working on more “traditional” humanitarian efforts focused on the basic delivery and distribution of aid.
Another less direct yet seemingly quite impactful limitation on open data that organizational representatives frequently cited was the concept of the privacy and ‘do no harm’ principles. Almost all surveyed organizations emphasized that any data sharing that they would consider must comply with the privacy and ‘do no harm’ standards to which they are committed. These are standards, it should be noted, that are universally acknowledged across the humanitarian community.
As it relates to affected communities, one representative of a medical humanitarian agency stated that there was “no obligation to provide data to the communities as a whole,” in part because “the [affected] community imperative never trumps the rights of an individual.” This very strict emphasis on protecting the privacy of an individual’s data was stressed most intensely by representatives of medical aid organizations, but the general idea seems to apply across the sector – and for good reason. After all, it would be irresponsible if, for instance, a humanitarian organization published data related to HIV/AIDS status that could eventually be traced back to an individual or household living in an area of intense stigmatization. Although perhaps such data could make medical officials’ lives easier at the macro-level, it could also threaten the safety of the individual or family in question.
But does this requirement for certain levels of privacy and a commitment to doing no harm mean that open data in the humanitarian community is stuck in the mud? The answer, of course, is not at all. However, it is clear in speaking with humanitarian representatives that solutions to opening data in a meaningful way to affected populations while still respecting privacy and safety of all involved will require creativity and innovation.
One frequent way of balancing the push for open data with these humanitarian requirements is through the aggregation of data. But this common tactic, it seems, is in no way a silver bullet. Even if the aggregated data is shared with affected populations – something that is certainly not a given – is it necessarily helpful to them? Does it really help an affected community in the Philippines to know the rate of a certain disease at the national level? Does it help farmers during a drought in South Sudan to know the total quantity of sorghum planted in a given state or district in years past? The answer, most likely, is no. So, how can these types of data be secured and shared with affected populations in a way that actually makes sense and makes a difference? Well, that is what seems to be one of the sector’s biggest open data challenges.
Undoubtedly humanitarian organizations need to first and foremost remember to “do no harm” – this principle exists for very important reasons – but our conversations over the past few weeks also suggest that the sector ought to be careful not to hide behind these principles as a way to avoid confronting the difficult questions of opening up data in a way that is both safe as well as meaningful and actionable. There are some organizations that are beginning to grapple with these questions, looking for ways to solve this admittedly very difficult riddle. But the solutions, to be most effective and found in the most efficient manner, will likely require a significantly more substantial pursuit of innovative methods of data collection, processing, and sharing within the sector than what, at least from our conversations, seems to exist so far.