I’ll be the first to say it. I wasn’t a big fan of Twitter during disasters from an operational standpoint. I’m still an optimistic critic but what I love about the field of both humanitarian research and practice is how humbled we can be by the nature of change and the opportunities for use in the road that lies ahead.
In the past sometimes I sat in the corner of the room during conferences and meetings reading papers about how Twitter was going to save the next disaster and the penultimate opportunity for big data. Coming from an operational standpoint I always had a difficult time piecing together how such big data noise would earn a place in operational decision making. I continue to be humbled. From a few experiences I’ve learned to understand that these needles in a haystack may have added value. But it’s about the design, persistence and understanding about practical use and research innovation that will help us find a better way.
In the end I’m a believer that the journey will be useful and hopefully with a longitudinal collection of experiences we will better understand the role of social media and Twitter during disasters; whether it is “useful” or not.
Those needles in a haystack
Part of the conversation that I’ve been listening to often explores big data and social media to find that one needle in the haystack that will have the opportunity to “save lives”. While this is aspirational, and maybe anecdotally true by some reports, I’ve consistently been more drawn to the potential wins that can be achieved among trusted networks of people around me during a crisis response. Maybe it’s not the individual voice that matters sometimes as the purported social media needle in the haystack but it is the trusted networks, related organizations, and people within this network that may be what we should also be exploring through a domain driven lens of analysis. So rather than focusing on that needle, let’s explore those inter-connected clusters.
Organizations vs Individuals in the Sea of Twitter Communications
A study by McCorriston et al in 2015 looked at Twitter accounts that represent individuals versus organizations. They found that although very few organizations (<10%) existed in the corpus they were highly connected. During humanitarian crisis we believe that trust networks in the information management world and operational decision-making are perceived as highly important.
Searching for Added value as a Humanitarian Responder
For many years, in different roles as a humanitarian responder I would frequently (albeit relatively briefly) think about Twitter information during the onset of a crisis. I would ask myself what could be gleaned from Twitter conversations among humanitarians for operational response.
This focus was likely (disclosing here) because I’m part of a humanitarian trusted network of responders. There’s often just too much information without a good, sustainable and feasible pathway to be able to glean information for decision-making in the roles that I had. While there are (like Humanity Road) and were many digital humanitarian organizations trying to parse out information on Twitter, I struggled for years to find added value in the response work that I was engaged in.
For many crises I would scan the curated twitter reports to find information relevant to my mission and work, and rarely found new pieces of information that I could operationalize. In many ways I was limited likely by time and my own limitations and creativity for outcome driven opportunity seeking 🙂 I was unable to find a linear or efficient pathway for use.
In the same year when McCorriston et al’s research team was exploring this topic, I was the coordinator for the crisis informatics program at NetHope. We were cycling from crisis to crisis trying to manage and make sense of different types of different and fragmented pieces of information for the operational purposes of providing connectivity to our consortium members (NGOs) and other partners (UN agencies, government offices) around the world.
Finding that cluster of needles during NetHope’s response to Vanuatu Cyclone PAM
That same moment of inquiry with Twitter returned during our activation to Cyclone Pam in Vanuatu. I remember our team was quickly looking at many information sources and curating as fast as we could with the purpose of quickly understanding what the NetHope membership were doing in this response and gleaning as quickly as possible in a remote setting, how we could support them- either in their sectoral area of response (ex. water, sanitation, health, shelter) and/or their need for connectivity. It was also helpful for us to know where they were going geographically in the response and what overall needs they were witnessing as they performed their assessments and commenced planning.
That was the context of the questions that I was tasked to operationalize. And it was at that moment when I quickly looked at Twitter feeds and Humanity Road’s sitreps; that I came to realize that for this crisis those information requirements seem to be present in these information feeds into ways that I had seen not before.
We quickly started looking more closely at Twitter reports that were curated by organizations like Humanity Road, further screening them for our membership. And began to following our member’s Twitter handles directly
our trusted network on Twitter.
Twitter handles at the organizational level.
So rather than searching for the sentiment of 1 million or so people talking on twitter during a crisis we decided to take on that tiny cluster of needles (our NetHope member twitter handles) and look into that tiny haystack for information that we were not finding elsewhere. This was
social media information + trusted network’s + tiny needles + content exploration for focused operationally driven situational awareness for NetHope response
And what we found was that– directly following our membership’s Twitter handles and also using curated Twitter social media feeds did have added value. While we were collecting and curating public sources (government situation report ultra situation reports) and a deluge of private emails and communications- this new social media information source was manageable and providing potentially new pieces of needed information.
We had a very ad hoc 101 level Twitter feed following our membership. What we found out was there were key pieces of information that told us where what and who among our membership was responding this crisis; at times faster and in more detail than our other sources. This information was added to our existing crisis informatics situation reports that were used to inform our field staff coordinating with the emergency telecom cluster in Vanuatu. The information from Twitter had added value – new information that was not found from other sources- and found itself in the final coordination maps which were shared among other partners. I was humbled and excited that Twitter had found a way into our work.
not because it was interesting and “new”
but because it proved it’s added value (both in content and curation speed) and earned its place in near real-time operation information management products for field coordination.