big data

Feb. 20, 2014 – Hilary Oliva Faxon

The planet loses forest at the rate of 50 soccer fields per minute. Deforestation can be devastating for biodiversity and forest-based livelihoods, but it’s not just a local problem: land use change contributes almost 20% of annual greenhouse gas emissions, driving global climate change. Unfortunately, large areas of forests are challenging to access, making them difficult to track and police. Global Forest Watch aims change that.

The platform, envisioned and managed by the World Resources Institute (WRI) with Google Earth Engine, the University of Maryland, and 40 other partners, documents near real time forest cover changes in over 200 countries on an interactive map. Zoom in to the region of your choice and scroll through time to see the when and where of forest gain and loss. The information has major implications for land tenure, REDD+ programs, protected area management, and indigenous and state natural resource claims.

Global Forest Watch harnesses amazing amounts of data – some from NASA, some that’s been languishing in USGS reels stored in South Dakota – and puts it online for free. It took 10,000 Google computers several days to upload a base map, said Google Earth Outreach and Earth Engine Engineering Manager Rebecca Moore, “This is when we [at Google] start drooling.”

Despite the data’s massive size, the resulting platform is both simple and accessible (play with it and see if you agree). Ease of use is perhaps the program’s greatest strength: “You can’t solve problems you can’t see. This gives the global effort a whole new set of tools going forward,” said USAID Administrator Rajiv Shah, linking the platform to Obama’s Climate Action Plan.

The launch featured an unusual, and remarkable, list of endorsements: the former president of Mexico, an Assistant Secretary of State, the head of USAID, the CEO of Unilever, a major Indonesian politician, Jane Goodall, an UN indigenous rights leader. The diverse group speaks to the tool’s range. With a few clicks, companies can monitor their supply chains, indigenous peoples can gauge encroachments on their land, and governments and NGOs can track illegal logging. Users can make their own maps, set an alert for changes in a specific area (for example a national park), post their own data and stories to the site, and potentially much, much more… WRI has left it up to users to apply the tool, though they have a strong supply of both good ideas and funding for on-the-ground partners.

A few things are especially interesting to me about the launch and the tool. Full disclosure: I’m currently on contract with a different program at WRI, but these observations and opinions are my own.

One: data is changing development. In academia, Yale’s Karen Seto use remote sensing imagery to model urbanization in India and China. Meanwhile LIDS’ Andrea Titus comments below about data debates around the post-2015 framework. Now big data is being harnessed to address global deforestation, a persistent and truly “wicked” environmental problem with complex ecological, social, and geopolitical implications. To see gray-clad federal employees cozy up with Bay Area software developers means big data has truly arrived.

Two: the private sector is critical. Even outside of Paul Polman’s short video speech, sustainable supply chains and the Consumer Goods Forum came up repeatedly. The private sector is embedded in the issue of global deforestation and has huge potential to limit the problem by committing to best practices. Global Forest Watch means companies can’t hide so easily, though whether that changes industry practice remains to be seen.

Three: green is global. Launching a tool that can be used by anybody, anywhere, for anything is a pretty remarkable idea. The openness of the platform and the “tell your own story” message make this a rather novel experiment, at least for the environmental movement. If Global Forest Watch can hit it half as big as Kickstarter, Instagram, and… dare we dream… Facebook, it will have made a big impact by raising the profile of forest issues worldwide.

So here’s the big question for hippies like me: can you hug a virtual tree?

[Image: globalforestwatch.org]

Feb. 14, 2014 – Andrea Titus

“Wanted: A New Data Revolution.” Such was the call issued by the United Nations High Level Panel on the Post-2015 Development Agenda last year, no doubt in response to the two defining features of the current data landscape in development policy. The first is the dubious accuracy of national and official statistics that currently serve as the foundation of evidence for decisionmaking. The second is the rapid rise of technologies, social media, and crowdsourcing capabilities that have the potential to unlock new sources of data that are, at least on face, appealingly transparent and participatory.

When the Millennium Development Goals (MDGs) were established at the start of the century, they placed a new focus on the importance of measurable development indicators on a global scale. From the start, however, there have been serious concerns about the quality of the data we use to assess progress towards the MDGs, much of which relies on infrequent censuses and extrapolation, filtered through national governments. As the World Bank notes just alongside its interactive data portal, “the quality of global data depends on how well these national systems perform,” and performance is uneven at best.

Enter “big data,” and the potential for new and nontraditional sources of information to transform the development landscape. Already, researchers are using mobile phone records to track the spread of malaria, or Twitter posts to detect epidemics. The use of big data is still at its infancy, but it shows so much promise that the development community responded enthusiastically to the UN High Level Panel’s call for a revolution with its own set of colorful military analogies. “As Napolean is alleged to have said, “[W]ar is 90 percent information,”” wrote Claire Melamed of the Overseas Development Institute (ODI), “and that applies to the war on poverty as much as to any military conflict.”

Like any other revolution, however, the “data revolution” is characterized by a number of sticky questions about governance. While some see big data as a means to increase transparency and accountability, and therefore as a means to check traditional power structures, others are worried that existing power dynamics will simply map onto a new scene. There are also questions about how to effectively engage private companies (like those providing mobile phone records in the example above) in the data revolution. How can we incentivize stakeholders, like private corporations, to participate and share data? Does unchecked data collection open the door to exploitation for profit? How can we appropriately account for concerns about information privacy? And how can we take advantage of participatory technologies in a way that is most fair and democratizing? Who is not at the table? What gets left out?

These are the kinds of questions that lead Paul Ladd, head of the UN Development Programme’s team coordinating the post-2015 agenda, to urge caution in embracing the rise of data in development. Big data has big potential, but the agenda is not particularly straightforward, nor is it particularly generalizable across contexts. “The data revolution isn’t going to be a big ship in the middle of the ocean with everyone following behind it,” he told SciDev.Net. “The data revolution in the end will be a coordinated flotilla of smaller ships on the sea who, after talking about it for a bit, have a sense of common purpose and are all going in the same direction.” One way we can start, Ladd argues, is by thinking about how new data sources can generate proxies that can then be used to check and verify what we think we already know. And this in of itself would be a huge boon to decisionmaking in the development space.

For more perspectives on the potentials and pitfalls of the data revolution in development, check out this blog series on post2015.org.

Photo credit: digitalart/FreeDigitalPhotos.net

Tag: big data

From Tree Huggers to Tree Modelers: environmentalists heed big data’s call

What kind of data revolution for development?