Eco-business: Opinion By Stefan Verhulst, January 4, 2017- The quintillions of bytes in data we generate every day offer potential economic, cultural and political benefits. Here's how private data can be used for social good through the concept of data responsibility, according to New York University's Stefaan Verhulst.
Women walk along a street filled with debris in Sankhu, a town in northwestern Nepal that was particularly affected by the major earthquake in 2015. The use of data allowed the government and aid organisations to optimise relief efforts. Image: United Nations Development Programme, CC BY-NC-ND 2.0
As climate change intensifies, catastrophic, record-setting natural disasters look increasingly like the “new normal” – from Hurricane Matthew killing at least 1,300 people in September to Typhoon Lionrock, the previous month, causing flooding that left 138 dead and more than 100,000 homeless in North Korea.
What steps can we take to limit the destruction caused by natural disasters? One possible answer is using data to improve relief operations.
Let’s look at the aftermath of the April 2015 Gorkha earthquake, the worst to hit Nepal in over 80 years. Nearly 9,000 people were killed, some 22,000 injured, hundreds of thousands were rendered homeless and entire villages were flattened.
Yet for all the destruction, the toll could have been far worse.
Without in any way minimising the horrible disaster that hit Nepal that day, I want to make the case that data — and, in particular, a new type of social responsibility — helped Nepal avoid a worse calamity. It may offer lessons for other disasters around the world.
In the wake of the Nepal disaster, a wide variety of actors – from government, civil society and the private sector alike – rushed in to address the humanitarian crisis. One notable player was Ncell, Nepal’s largest mobile network operator. Shortly after the earthquake, Ncell decided to share its mobile data (in an aggregated, de-identified way) with the the non-profit Swedish organisation, Flowminder.
+ claudio, Author provided
Flowminder then used this data to map population movements around the country, and these real-time maps allowed the government and humanitarian organisations to better target aid and relief, maximising the impact of their efforts. The initiative has been widely lauded as a model for data collaboration.
And Flowminder even won an award for “Mobile in Emergency or Humanitarian Situations” at the 2016 GSM Association’s Global Mobile Awards in Barcelona.
To me, the most striking part of the Flowminder-Ncell initiative is the way that data was used. In particular, how data originally collected for private purposes was exchanged for public ends: an act of data responsibility.
Data responsibility – and corporate data sharing – is an emerging concept, still in development. But it is becoming increasingly apparent that it can play a central role in fostering a variety of public ends, including in the way we respond to natural and other disasters.
Data responsibility could also play a central role in fostering the 2015-2030 Sustainable Development Goals. As Jeffrey Sachs has noted: “The data revolution can drive a sustainable development revolution, and accelerate progress toward ending poverty, promoting social inclusion, and protecting the environment.”
According to recent estimates, 2.5 quintillion bytes of data are created every day; nine-tenths of the data we have today was created in just the two years preceding 2015.
This data explosion has generated considerable enthusiasm for its potential economic, cultural and political benefits. The Economist has written of turning “dross into gold” by mining the huge streams of “data exhaust” often inadvertently left behind by always-connected users of social networks and mobile devices.
What is less discussed, however, is that most data remains locked up and proprietary, the private property of companies, governments and other organisations. This limits its public benefits.
Shifting from a habit of data shielding to data sharing and from traditional policy making to data-driven governance will require a cultural transformation in the way companies, governments and other actors treat their data.
Data responsibility can help organisations break down these private barriers and share their proprietary data for the public good. In the case of the private sector, in particular, it represents a type of corporate social responsibility for the 21st century.
Today, data responsibility remains relatively uncommon. Ncell, in Nepal, is one of relatively few corporations that have opened up their vast troves of data.
But there are a few encouraging signs. In Jakarta, for instance, Twitter shared some of its data with Australian researchers, who used it to create the website PetaJakarta.org. It provided real-time intelligence on flooding, enabling far better assessment and improving management, particularly during monsoon season.
In Senegal, the Orange Group initiated the Data for Development challenge and shared its data with different teams of researchers in order to identify patterns and solutions that can improve health, agriculture, urban planning, energy and national statistics.
The winning team used mobile phone data to provide an accurate proxy of energy needs, allowing for bottom-up solutions to fluctuating energy demands.
Such examples show us that data can improve and even save lives. But in order to fully harness the potential of data, three conditions must be fulfilled. They comprise the three pillars of data responsibility.
This is perhaps the most evident duty: to share private data when it’s clear that it will serve the public good. Secondary use is not always popular among data holders (often for good reasons) but when done correctly, data sharing can have powerful social benefits, as illustrated above.
Sharing does involve risks, notably to privacy, security and other individual rights. So it is imperative that organisations share responsibly, with every effort to protect both the data itself and the individuals who have surrendered their data (even if often unwittingly).
The consequences of failing to protect data have now been well-documented. The most obvious problems occur when data is not properly anonymised before it is shared, or when de-anonymised data otherwise leaks into the public domain.
Ostensibly anonymised data may itself also be susceptible to de-anonymisation, wherein information released for the public good ends up causing individual harm.
For example, in New York City in 2013, the Taxi and Limousine Commission, responding to a public request, released supposedly anonymised information data on pickup and drop-off times, locations, fares and tip amounts, collected from various taxi companies and ride-sharing firms. But within days, several civic hacker groups had managed to identify relevant taxi licenses and medallion numbers.
The consequences were worrisome and potentially rights-violating: the data could be used to calculate a driver’s annual income, for example, and to identify consumer travel and spending habits, including details on several celebrities, which raised their risk of stalking.
Thus the good intentions guiding data releases must be accompanied by a powerful sense of responsibility at every stage of the information chain, from data collection, processing and analysing to sharing and use.
For released data to serve the public good, officials and others must also adopt policies and interventions based on insights gained from its release. Without action, the potential remains just that — potential.
This duty to act is particularly evident in the struggle against corruption. Around the world, datasets released by governments (and other organisations and individuals) have played a powerful role revealing corruption and increasing transparency.
Brazil’s Transparency Portal, for instance, created in 2004 by the Office of the Comptroller General to increase fiscal transparency by sharing government budget data, is now one of the country’s primary tools to identify and document corruption, registering an average of 900,000 unique visitors every month.
In Mexico, the online platform Mejora Tu Escuela provides citizens with information about school performance, helping parents choose the highest-quality education for their children and get involved in their schooling.
For school administrators, policymakers and NGOs, on the other hand, the platform provides a way to identify corruption in the form of “ghost teachers” on government payroll and teachers who are paid apparently outsized salaries.
Still, to translate insights into impact, action is needed. And this often relies on vast and difficult changes in the face of vested interests and institutional obstacles.
The difficulty of translating insights into results point to some of the larger social, political and institutional shifts that will be required to achieve the vision of true data responsibility.
Data responsibility requires an often unfamiliar commitment to values of transparency and accountability. Shifting from a habit of data shielding to data sharing and from traditional policy making to data-driven governance will require a cultural transformation in the way companies, governments and other actors treat their data.
The following three ways should be considered to bring about the required transformation in the short term.
First, public and private data holders should issue a public commitment (or pledge) to data responsibility, so that it becomes the norm within organisations, rather than the exception.
Second, the position of “data stewards” should be created within public and private organisations. They will act as change agents and determine what and when to share, how to protect, and how to act on available data.
Finally, we need a movement: it is time to expand the right-to-information community to include a demand for “data responsibility” to help improve lives – including ours.
A version of the article was presented by the author at TEDx MidAtlantic on October 20-21 in Washington DC.