Not long time ago, actually just in June, I was at a bar in Boston discussing with two friends of mine, Nona Lambert and Jaro Valuch, from the Ushahidi Haiti Project, about emergencies and the use of the Ushahidi platform. We were comparing our experiences in Chile and Haiti and trying to understand how we could face a new emergency knowing what the biggest challenges could be. Our first main concern was how to be able to increase the use of crowdsourcing for tasks related to the management of the Ushahidi platform in emergencies situation.
I am a very big fan of decentralized non-hierarchical networks as opposed to structured hierarchical systems, but there is problem that we faced and that still will be faced by new Ushahidi instances in emergency situations: how can we reach in little time the people out there, that have the skills and the time and the willingness to help? And after we reach them how do we organize their work in order not the waste our time and end up with overlapping, or waste of incredibly useful resources?
My solution at this point is that of course the crowd manages itself. Ok. But how? Here is where technology comes in.
Pakistan is the new humanitarian crisis, and of course a Ushahidi platform is already there, with a short code and a team working hard to make this project useful. New emergency, old challenge: how to reach to people, inside and outside Pakistan, that are willing to help, from Pakistani people living in less affected areas that wants to help their conationals, to NGOs, CBOs, Universities, media local authorities and other organizations inside and outside Pakistan that can provide assistance and help. And ones reached to them, how to organize their work?
One example from Haiti was the 4636.com website, where people could subscribe and translate messages from Creole or French into English: this was a perfect example of organization of the crowd by itself with the use of technology. You could just subscribe and translate, no overlapping, no waste of time or energy: ones you were taking your messages the program was preventing others from doing the same, and everybody could work without worrying of having to coordinate with others.
For the Pakistan instance, my friend Patrick Meier is suggesting to go further: we could have a system like 4636 also for tagging and geo-location of information. In this case, he suggests, to have triangulation to verify the geo-location by creating an algorithm that will send a message to the Ushahidi platform only after three people have tagged it in the same area, let’s say 500m range. If the three locations are slightly different a line in between the three points could be drawn on the Ushahidi map to show the area.
I think this is a great idea, but we still have a problem to be solved that the Ushahidi Pakistan team is facing, as we were facing: how do we reach to the people, to the crowd? They are out there, and sometimes is not easy for them to get to know that there is way to help, from their home, with their skills and capacities.And the same is valid about how to have the responders using the platform, and in this way assure a timely response to the needs.
When working with the Ushahidi platform three different skills can make a huge difference: the translation, or just the comprehension of a language, the mapping, and the ability to answer to the needs. The mapping, I will say, is the easiest one: it is really just a matter of practice, you have to be careful of course, but the more you map, the more you learn how to map, where to find the location if not on a map, how to get the right information to find the right place and so on. Tools like OpenStreet Map, Google Maps and Google Earth are increasingly easier to use.
But the other two skills are skills that you have or not: a language, you speak it or not, no way you learn in time for an emergency, which is normally max a week. The ability to respond is normally related to your job and to your network, again, you have it or not.
Now the idea me, Jaro and Nona had was to use a tool that already collects this information from the people using it: Facebook.
This is how a system like this could function. Your Facebook page is normally set on your language. If you speak more than one language, there could be a function on your page were you can write which languages you speak, or even easier, there could be an automatic system to detect the languages used on your Facebook wall. On the other side in your Facebook page there is also a section where you list your jobs, present and past.
Now, lets’ say that Facebook creates an option for its users to allow them to choose if they agree on receiving a call for action in case of emergency. It is pretty easy to decide if something is an emergency or not: we all agree that Pakistan is an emergency, as we all agreed that Haiti was and Chile too.
In case of an emergency like Pakistan, the Ushahidi platform team could just use a system by which they insert the main characteristic of the people they want to reach, in this case could be: 1- Speaking English, Urdu and/or Pashto 2- Living in Pakistan or having Pakistani origin 3- Working in Pakistan (and would leave this general).
By choosing the criteria a program would just match the characteristic together and send a message to all the people with those characteristic using Facebook, directing them to the page where they can find the instruction to help. In this way the people sending the message, in this case the Ushahidi platform team, don’t have to see the information of the people directly (who may be concerned about having their info accessible to others that they don’t know) but just to send the message out.
This could be also done with the mapping system, but having people using systems like 4636.com to map and having a score assigned to them, as Patrick suggested in his last post.
This system could be used to create a top 100 or top 1000 list, with people having their score published on their Facebook page, but also becoming “crisis mappers” after a defined number of right locations find, in order to have them as “accredited mappers” that don’t require triangulation on their reports. A random verification system will still be needed according to me, but it could be just done by the administrators of the platform.
This system will not of course prevent others from joining the effort, but will create an automatic roster of people, with specific skills that could be mobilized immediately in case of emergency and decrease the waste in time and energy in the overall mobilization of the crowd. In addition to that people participating in the effort will advertise their involvement, in this way recruiting others, in their network and circle of friends, in joining the effort.
I think that Facebook could be one tool, but lots of others are available out there, where the information and the skills of people are already collected and available, like LinkedIn, Twitter and so on. The idea is that everybody can join and work on it, but also that people that have particular skills or networks, particularly useful in a certain situation, could be reached more easily skewing the balance in between time and accuracy, not only of information but also of mobilization.