Haiti study: Mass mobile phone tracking can be laudable
A new study uses the movements of mobile phones during the Haiti earthquake, and cholera epidemic, to accurately show where people went during the disasters, and where help should be delivered.
Studying location data stored by Haiti’s biggest network operator, Digicel, Swedish boffins got more accurate estimates of population movements over the period of both earthquake and epidemic than rescue workers on the ground, demonstrating the value of anonymous real-time tracking during national disasters.
The research is published in the peer-reviewed PLoS Medicine journal, and provides a detailed breakdown of the data gathered by the Swedish researchers (one of whom was American, to be fair).
It might seem obvious that the location of every mobile phone would tell you where the population was, but proving that required extensive analysis of the data, as well as adjustments for the penetration of mobile telephones and the demographic differences involved.
The first problem is that not everyone in Haiti has a mobile. The earthquake study tracked 1.9 million SIMs, having discarded numbers that had not made a call in the preceding month (to exclude rescue workers) and those that did not make a call afterwards (euphemistically described as “lost”).
Around 200,000 of those SIMs exited Port-au-Prince following the earthquake, which, given the mobile penetration of just under one-third, multiplies up to 630,000 people leaving the capital.
That tallies well with official UN figures, compiled through questionnaires following the disaster, but diverges from estimates made at the time (which were based on a counting of buses on the roads exiting the town).
Even more interesting were the results from the cholera epidemic, which were compiled in less than 12 hours and demonstrated that such a process could provide real-time advice to healthcare workers and governments in containing and treating, as well as tracking, outbreaks.
The researchers suggest that localised SMS could have been used to advise people within affected areas, or to discourage them from travelling elsewhere, as well as to let the authorities know where problems might surface next.
That does, of course, require close integration between the operators’ systems and those of the government, which will make many people uncomfortable. Real-time tracking, even by cell site, would have been valuable (for example) to the police during the recent UK riots. That data is already being used retrospectively to work out where people were, and with better integration it could show where people are.
The Chinese government recently launched a research project working out how such data could be effectively used in urban planning and traffic management, but if it were to be available in real-time it could also be used to police demonstrations, football matches and all sorts.
This needn’t be an invasion of privacy – anonymous data is still valuable – but we have to decide, as a society, if we think that allowing governments access to the location of their citizens is a risk worth taking in exchange for the benefits it gives. ®