eq.org.nz - The Power of Ushahidi
When the Christchurch earthquake struck Tim McNamara and many others didn't just wait for someone else to step in and do things, they took advantage of ICT and just did it. Here's Tim's story, as told by himself for NZCS Newsline.
One of the biggest problems in the immediate aftermath of the Christchurch earthquake was a lack of information. For example, of the petrol stations that were operating many only had diesel left. Others had tanks reserved for emergency services vehicles. Even in places that were lightly affected, supermarkets had changed their hours.
Official information sources and processes to gather and disseminate this information were patchy. A Wordpress instance was established by council and Gmail accounts were created on the fly. There was only a small, overworked team processing the information and publishing it. They were housed in the Art Gallery, situated in Christchurch's Red Zone. This area was swamped with responders using mobile Internet.
The Wordpress instance reached capacity problems, probably due to the fact that it was an off-the-rack package from Automattic. At times the city council requested that viewers visit the site via a third-party web proxy. Having local servers didn't always help; being powered by generators, servers experienced temporary outages.
NZ Tech Community Response
The New Zealand tech community responded to the Christchurch quake by harnessing crowdsourcing. We created a system that parsed hundreds of thousands of tweets, emails, SMS messages and web form submissions. This system, staffed by volunteers, distributed information through media of all kinds.
We were even on page 700 of Teletext, I kid you not.
Our focus was on providing information that was relevant to people using information. A content management system is useful for the people producing the content. However, it's impossible to ask for a single list that contains status information for the local school, coffee shop, pharmacy and water tanker from a blog. Instead, an article tends to be created for each type of information. This means that people under stress are required to read through lists of each of those categories themselves.
The Christchurch Recovery Map could be updated by people with an interest in keeping their own information current. Retail banks were updating their own ATM information many times a day as cash levels changed.
Fisher & Paykel appliances decided to use eq.org.nz to promote their free laundry services. It's unlikely that the nearest open cafe will ever appear on a civil defence publication. However, these community assets are important too. Proving a neutral space for crowdsourced information meant that we could combine official information about essential services with less critical information, such as locations of free BBQs that were established by neighbours helping each other out.
What we used
The technology solution itself is called Ushahidi, a Swahili for testimony. It was established to monitor election violence in Kenya but he project then expanded to become a system that enabled crowdsourced information to be gathered and published in any crisis.
Ushahidi works by using human curators behind the scenes. Every incoming message that contains a fact and a location is categorised and plotted onto a map.
Our particular instance was able to verify, categorise and publish new information within a five second to five minute window of receipt.
All of our data is then published via an open web application programming interface that allows third parties to combine our information with their own. A good example of this was Environment Cantebury's own map viewer, http://eqviewer.co.nz.
The eq.org.nz also domain became a home for several other projects. This includes GIS data storage, the nominated URL for Google's Person Finder, peer-to-peer business support, a community projects space, printable maps and a mobile interface.
We also provided our free SMS shortcode to the Student Volunteer Army for its use. This meant that the SVA could receive requests for help from people without power.
How it happened
Here are some of the important events during the course of the project. I haven't put specific dates in, because the exact order is difficult to determine:
- CrisisCommons International activates its Standby Volunteer Taskforce and an Ushahidi instance on crowdmap.com. Two other sites are established, including one by stuff.co.nz.
- Koordinates funds a full powered Amazon EC2 instance to replace the turn-key crowdmap.com instance suffering performance issues.
- All three telcos approve zero-rating of the SMS shortcode; agree to coordinate their social media efforts to support the project
- Catalyst IT fund a dozen of its staff to eliminate initial backlog, provides for SMS shortcode to be free
- Victoria University of Wellington provides space for volunteer training (the next day, my old rescue team NZ-RT7 is deployed to Christchurch)
- stuff.co.nz & nzherald.co.nz embed the map into their sites; stuff.co.nz agree to replace their own Ushahidi instance with ours
- Google amends Google Search to have the map as the first result for when searching for "christchurch"
- Optimal Usability agrees for its staff and venue to host weekend training of volunteers
- Trademe links to our site from its earthquake resources page
- Fair Go features http://eq.org.nz on its Christchurch Earthquake special
- Student Volunteer Army partnership formed; reports that site is heavily used by community groups for needs assessment
- Nigel McNie fixes Red Cross New Zealand's donations website
- The New Zealand Geospatial Office hosts a briefing for officials and media. Several government departments attend. Officials opt-in to be volunteers themselves. The project appears on TV3 News
- Department of Internal Affairs and other public sector bodies promote the site heavily via their Twitter feeds
- The map reaches 100,000 visits
- InternetNZ funds me to fly to Christchurch to make local connections and provide advice about how the organisation can help
- Custom printable maps from our data available, updated at 10m intervals
- Local adoption was not forthcoming; normal information channels resumed; site usage experiences decay curve
- Decision made to announce "Mission Accomplished" and allow volunteers to rest
The utterly amazing people
Above all else, the Christchurch Recovery Map was made possible by the efforts of several amazing people. I'm reluctant to name anyone, because it necessarily leaves people out.
However, some especially noteworthy project members include:
- Justine Sanderson & Robyn Gallagher for their amazing dedication by each filtering tens of thousands of messages
- Rob Coup, Nigel McNie, Richard Clark, Brenda Wallace & Sam Minee for providing tech skills that mere mortals do not possess
- Josh Forde
- Tim Kong, Melia Meggs & Demelza Wood for being terrific with volunteers
- Anthony Baxter, Nóirín Shirley & Penny Leach for coordinating international support
One of the things that I think made that possible was that, at least initially, we had a high barrier to entry for new helpers. Things were chaotic and processes hadn't been firmly established. Volunteers needed to be self-directed and confident enough to ask for help.
High social capital
Social capital is a phrase used in academia to describe the communities' connectedness. A community with high social capital has many people deeply engaged in sports clubs, churches, and that sort of thing. Members within these communities are very able to call upon the resources of others during adverse events. New Zealand's tech scene has high social capital.
One of the biggest success factors were individuals such as Nat Torkington, Tash Hall Lampard and Sam McDowell. Nat had access to the Kiwi Foo alumni. Tash was able to join dots on the Webstock mailing lists. Sam worked with the NZ Bankers Association to have banks contribute their own status information. Without these connections, the fabulous people that provided support would never have emerged.
One of the things that struck me was the breadth of skill that software developers who are familiar with open source development posses. Developers used to developing on a Microsoft platform were unable to install a local instance of a web server, get to grips with version control software or communicate effectively with tools used for synchronous remote communication, such as IRC.
In the end, several of the developers who came from a Microsoft/.Net platform went away disheartened.
The open source model was critical for the success of the project. There are technical as well as personal reasons for this. From a technical perspective, the open source community shares several practices and norms for effective remote communication and collaboration. This meant that it was simple to manage a software project that had developers working in multiple time zones in a very constrained timeframe.
However, a larger part of the success was due to the ability for multiple people and organisations. There are few projects in the world supported by Google's Chief of Engineering as well as a high school students, simultaneously.
While Craig Neville-Manning may not have written any code for the project himself, he certainly approved the use of Google's resources to move the project along. Without the open source mentality, I don't think that it would have been possible for the Apache Foundation, Catalyst IT and the CTO of SilverStripe to work together without prior arrangement. As it stood, everyone could opt-in to the level that they were comfortable with. Everyone participating in the project understood that they could work on bug tickets that suited their skill level.
That spirit of collaboration has enabled the efforts of eq.org.nz to support the people of Libya and Japan. Many of the improvements made by our team have already been incorporated in deployments to those areas. This includes everything from behind the scenes work, such as optimising clustering algorithms, to front-end usability changes made.
Moreover, the project was completely vendor-neutral. As far as I could tell, we were the only website providing service to the community without corporate logos. Neutrality meant that we a had high levels of trust and meant that it was safe for competitors to cooperate.
Our platform was neutral to non-code contributions also. This meant that we could provide non-essential community assets like bouncy castles in parks alongside critical information like water distribution points. The neutrality meant that while individual telcos had incentives to provide the locations of only their wifi hotspots, our incentive was to publish all.
In the end, most information in the system came from organisations reporting on their own facilities. They probably used Twitter as a medium to distribute that message. This message was picked up by our volunteers and plotted. Over time, businesses and the public sector began to input information directly. This distributes the burden for the information system to the organisations that have an incentive to keep it right, such as local business owners. Without neutral ground, the site would not have prospered as it had.
Many ideas have sprung up over the last few weeks about how to best utilise the spirit and calibre of New Zealand's technology industry. Here are a few that I personally think have a shot at doing well:
- One Christchurch, a project conceived by the Student Volunteer Army's parent body, to coordinate the efforts of community groups in the city
- creating a turn-key system that can be used for future disasters to receive and disseminate information from every source in a scalable, trust-worthy way. In effect, could enable any business to report to its customers and staff what its status is in a very easy-to-use way. This could reduce call centre loads, staff stress levels and lead to
- harness the ubiquity and resilience of mobile to support disaster reconnaissance and rapid needs assessments
Feel free to get in touch if you would be interested in learning more about these ideas.
You must be logged in in order to post comments. Log In