"There is complete chaos here. Nobody knows what is going on."
Those were the first words my Nepali colleague Jigmy Lama said over the phone when I called him a few minutes after we heard about the earthquake. We had felt the tremors in our Delhi apartment and rushed out of our building. Slowly, news trickled in through Twitter about the epicentre being in Nepal. By the time my call to Jigmy and other friends in Nepal went through, shock, unlike dust from collapsed buildings, was slowly settling in.
However, post two earthquakes, the pattern within that terrible chaos was revealing itself. On the broad canvas of social media, every dot of hope, despair and cooperation was getting geo-tagged.
" There is very little capacity in government and relief agencies to absorb the information being put together online and act upon it."
The huge amount of digital data that continues to grow after the earthquake could not have been possible without the spurt in growth of Nepal's telecom networks and cellular internet services. Though they are not as good as some people's expectations, the quality and affordability of telecom services in a politically hamstrung country is praiseworthy. Through 3G services and smartphones, huge volumes of photos and experiences were shared on Twitter and Facebook after the earthquake. It brought the horror home, and helped grab international attention. Piggybacking on this trend, crowd mapping and reporting arrived on the scene. With an aim to turn each social media post into rich data, it has set new benchmarks. This is where the earthquake in Nepal may become a turning point for disaster reporting.
Leveraging on local developers and whatever Internet was available, a number of web-based platforms were set up in no time. The most talked about currently is that of the Kathmandu Living Labs, which is mapping reports from affected areas, and connecting them with relief workers. Their platform has helped in lowering the information asymmetry that typically plagues a post disaster scenario. The second initiative, a joint one between Kathmandu Living Labs and the Humanitarian OpenStreetMap Team, is providing offline maps for relief workers to locate places, as well as providing guidance on how to use OSM on Android phones to map damage.
Around 2200 volunteers across the world were mobilised to map road networks, damaged houses, and even helicopter landing sites for relief operations. The People's Association for Himalayan Area Research (PAHAR) has provided downloadable online high-resolution toposheets. Micromappers have created online crises maps by analysing 35,000 images and over 7,000 tweets. This has enabled them to categorise a specific structure/point as either severely or mildly damaged. They are sharing their data feed with Kathmandu Living Labs. Another company, Digital Globe has activated its crowd-mapping platform, tomnod, which allows any user to plot damaged roads, buildings, etc. on high-resolution satellite layers.
In order to crowd source verification of images and their locations, along with local updates, Verily has also been activated by Crisis Mappers. A group called Map Action has also put up downloadable maps on actionable items, such as identifying priority districts for medical assistance. Jostling within this for some elbow space is a local crisis-reporting platform Bhukampa.
"[R]elief workers are moving from one platform to another like pinballs... many often do not know what they are looking for."
Crowd mapping is currently a crowded space in Nepal. However, producing maps and data is just one part of the story. Getting them consumed and acted upon for relief, rescue and coordination is another. It has been relatively easier to get thousands of volunteers across the world to work together on maps than getting 10 organisations to work together on the ground to save lives.
This is perhaps one big lesson emerging from Nepal's post earthquake scenario. There is very little capacity in government and relief agencies to absorb the information being put together online and act upon it. Data mediators, who interpret the data to relief workers, and then speak to mappers and data scientists to communicate actual need on the ground, are too few and perhaps as much a need of the hour as other specialised services.
Secondly, there is a surfeit of content. In one way, it's a good thing to have as many data platforms as possible. On the other hand, given the chaotic environment, it is too much to expect organisations to sift through multiple platforms and glean the information they need. Currently, relief workers are moving from one platform to another like pinballs. This problem is amplified by the fact that many often do not know what they are looking for.
Finally, though Nepal is flooded with 200 relief agencies, there is very little public information about their current location, activities and challenges. This is further compounded by the opaqueness that shrouds government relief efforts. There are surplus supplies in some areas while others get ignored. The fundamental question of disaster relief coordination -- "who is doing what and where" -- remains unanswered. Nepal's telecom and cellular Internet services have proved to be incredibly resilient. Hence relief agencies should use the Internet a bit more for coordinating their work. Ideally if every organisation tweets details of their supplies every time they are sent out, and confirms delivery with a tweet or an SMS, it can provide valuable information to other relief actors. A local youth group, AYON Nepal, is doing this very effectively and its time the government and international agencies picked it up.
"Government and relief agencies now need to engage with Nepal's resilient mobile Internet services and crowd mapping initiatives... Time is of essence, and unlike bandwidth, it is definitely running out."
This perhaps is another lesson learnt. There is no single open-source disaster management platform where users can plug in SMS, tweets, updates from multiple platforms to generate real-time demand and supply maps. This can be a game changer in managing disaster relief in the future, and bring in greater transparency. Although Twitter has become a default choice, we must remember that it's very hard to categorise, archive and make content searchable on this platform. Thus it is unwieldy for relief coordination purposes. So the sooner we get to develop a platform that does not have such pitfalls the better.
Government and relief agencies now need to engage with Nepal's resilient mobile Internet services and crowd mapping initiatives to coordinate between them and be more transparent. Time is of essence, and unlike bandwidth, it is definitely running out.