Thursday, June 6, 2013

Finding a definition and purpose for sensor journalism at the Tow Center

#Towsense presentation on mapping mangroves
by Aaron Huslage, photo by Moshin Ali (@moshin)
Before I left for the Tow Center's sensor journalism workshop at Columbia University last weekend, my wife and I hosted some friends for a spinach lasagna dinner. On the stack of books my wife was researching for her dissertation on toxins in science fiction, sat my latest obsession, the DustDuino sensor node.

While we ate, the node's tiny LED lights blinked away as it took particulate matter readings every 30 seconds. A friend pointed out the interesting juxtaposition of the pollution monitor siting on a copy of Rachel Carson's Silent Spring. Our friend's husband, an entomologist, asked what it was all about.

"Part of the idea is to have people make these all over the country, especially places with bad air," I said.

"Interesting," he said. "You know... that doesn't sound much like journalism," he said. "It sounds like research."

I thought about it for a moment, and took a sip of the lemon-and-bourbon cocktail my wife prepared. I didn't have a good answer.

"Journalists are kind of unemployed at the moment, so we're looking for other things to do," I replied.

For me at least, the Tow Center's workshop helped find an answer to that question, and provide a deffinition and goal for sensor journalism. About 50 folks with backgrounds in journalism, science, architecture, community informatics, and computer technology came to the Tow Center's first sensor journalism workshop on June 1-2.

I owe a big debt to the organizers of this event: Emily Bell, the Tow Center director; Fergus Pitt, Tow research fellow; Taylor Owen, Tow research director; along with Laura Kurgan, director of the spatial design lab at Columbia; and Chris Van Der Walt and Sara Jayne Farmer of Change Assembly, Inc.


 Everything Old is New Again


Brainstorming story ideas at the #towsense workshop. Photo by Taylor Owen (@taylor_owen). I'm pictured just left of center, nearest the blue DustDuino cube.
In one of the first presentations of the weekend, Mark Hansen of The Brown Institute for Media Innovation pointed out that sensor journalism builds on some very old concepts. Operation Igloo White, for example, was a US effort during the Vietnam war to drop acoustic- and urine-detecting sensors along the Ho Chi Minh trail. Readings from the network of 20,000 sensors in Vietnam allowed American bombers to hone in on Viet Kong trucks and tanks moving along the trail.

Of course, that's not sensor journalism, but rather a military operation. But the goals of sensor journalism projects share much with Igloo White: deploy sensors, collect data, and analyze that data for some greater meaning.

It was argued at the workshop that sensor journalism might actually be quite old. Cameras, after all, are sensors. They remotely sense the light levels of the environment, and store that data in the form of an image (either in analog film grain or digital pixels). Our interpretation of the image is data analysis.

David Eaves, advisor to the Mayor of Vancouver on Open Government and Open Data, argued on his website how sensor journalism is an expected part of the normal, everyday newscast. It may even take up as 5-15% of the news. It's just called "the weather."

"The meteorological group is a part of the news media organization that is completely reliant on sensors to provide it with information which it must analyze and turn into relevant information for its audience," he wrote. "And it is a very specific piece of knowledge that matters to the audience. They are not asking for how the weather came about, but merely and accurate prediction of what the weather will be. For good or (as I feel) for ill, there is not a lot of discussions about climate change on the 6 o’clock news weather report."

Eaves pointed out that traffic reports, which rely on various sensors, also qualify for that category of sensor journalism. But because it's mundane, not flashy, and also not inherently risky (more on that concept in a moment), the initial reaction of many would-be sensor journalists might be to ignore it. It's important to learn from that sector's experts, to internalize their methodology and build on it.

So sensor journalism probably is not a new concept. The rough deffinition of sensor journalism I gathered was: sensor journalism is a subfield of journalism that relies on sensing devices, and the data obtained from those devices, to arrive at an evidence-based conclusion in the course of a journalism investigation.

Because drones essentially are just unmanned vehicles on which to place sensors, drone journalism (especially the kind that relies on geospatial data) may very well fall into the wheelhouse of sensor journalism.

Yes, there is research involved. But the goal is much broader, and the end product looks much different.

My Sensors Indicate Danger

Photo of the DustDuino sensor node. Photo by Jeff Ginger of the CU-Community Fab Lab.

Speaking of experts, sensor journalists should collaborate with the experts who've been monitoring things with sensors as part of their research. We need to learn about best practices of establishing networks to obtain the most reliable data, and perhaps more importantly, to have a trained eye look over our results to determine if we can be confident in our conclusions.

It isn't just good practice, but ethical given the implications of a false positive, and even legally prudent. The workshop was very lucky to mine the expertise of Joanne Irene Gabrynowicz, the director of the National Center for Remote Sensing, Air, and Space Law at the University of Mississippi School of Law, during a panel discussion on ethics, law and politics.

In her opinion, the best protection against a journalist or publication from legal action resulting from a sensor journalism investigation is to rely on an expert to review the data and analysis. Journalists using sensing devices might be able to establish a solid defense in the courts, so long as those journalists based their investigations on the advice of respected professionals who adopted the best practices of their chosen field.

I don't consider myself a techno-optimist, i.e., a person who thinks every advancement in technology unequivocally improves society. Generally, I view the impact most advancements as simply shifting a line between two or more trade-offs. My attention span has diminished greatly since I got my first smartphone, and it's harder than ever to "turn off" work when out of the office. Say nothing of what automation and globalization has done to the blue-collar workforce.

Still, I can get giddy when discussing whiz-bang "disruptive" technologies like drones and crowdsourced sensor networks, such that my optimism takes some time to be dulled. The ethics, law and politics panel was helpful for identifying the caveats which should slightly curb my enthusiasm.

Sensor journalism relies heavily on data journalism and "big data" practices to make use of all the data that's being sensed. It follows, then, that the same ethical considerations for big data also apply to sensor journalism. Kord Davis, Co-author of Ethics of Big Data, made the point that there are "unknown unknowns" in the world of data collection. The data we obtain from a sensor might seem harmless one day, until someone figures out a way to triangulate that data with other data in order to fingerprint and track a person.

To use the immortal words of Google's Eric Schmidt, “We know where you are. We know where you’ve been. We can more or less know what you’re thinking about.”

Davis' solution was to destroy the data, so that it couldn't live long enough to be used for some kind of nefarious purpose. I think the idea deserves thought. How much of our data should we detail? How long should we hold on to the data? What should we keep, and what should we destroy?

Even so, Davis had some optimism about the future. It took Upton Sinclair's The Jungle to change the meat packing industry. It took Carson's Silent Spring for environmentalism to take off in America. Davis said he was still waiting for the equivalent exposé for the age of sensors and big data.

There was some talk of the IRB (Institutional Review Board) process for approving research at universities as a method of identifying and resolving ethical issues in sensor journalism investigations. Matt Waite, the founder of The University of Nebraska's Drone Journalism Lab, pointed out that IRBs can be an onerous process that may intrude into prior restraint. However, he said, there may be merit in an organized, methodical review process.

Sensor Journalism as a Product of Post-Industrial Journalism


Since the people we used to rely on as sources now publish eyewitness accounts on their own, our status as gatekeepers has diminished. Journalists had to pick higher-hanging fruit to feed themselves.

Despite the mistakes of Reddit (perhaps none so catastrophic as the crowdsourced attempt to finger the person responsible for the Boston Marathon bombing), it is providing a service in service-deprived areas. In the college town where I live, the UIUC SubReddit routinely provides the scoops and contextualization that local media just can't. Specifically, I'm thinking of a recent car fire, and a suicide in which local Redditors crowdsourced mental health resources.

Emily Bell, the Tow Center's director, along with C.W. Anderson and Clay Shirky, wrote and spoke about the fact that we're beginning a new age of journalism, called "post-industrial journalism." News will no longer be manufactured in large newsrooms, partly because the advertisement subsidy has been undercut by the internet.

Breaking news is now covered by the crowd. There is, as Bell said, no longer "the press."

"The speed of uploading images and video, the quality and length of video on camera phones, the ability to stream live events from a phone without a battery of attendant satellite trucks, and the frictionless sharing of all material through social recommendation transform our expectation and experience of news," Bell wrote for The Guardian, after the Woolwich killing.

"We still know very little about the planning and motivation for the attacks in Woolwich, but we know the tools of recording and dissemination are leading us into a world of streamed events and atrocity which will find us, unfiltered, through the phones in our pockets."

Even though I wasn't able to articulate at the dinner party, I knew in my gut that sensor journalism could be part of a larger plan to make us useful again. At the end of my piece on the Page One documentary of the New York Times, I made a comparison between a government intelligence and a news agency.

There are big differences, no doubt, but my point was that journalists should be functioning more like an intelligence agency in the sense of providing greater context with a greater understanding and access to scientific approaches. The biggest difference in the product of the intelligence agency and a news agency is the packaging:
The public should be provided these same services. The world is awash in data. What it lacks is understanding. The fourth estate is that it is positioned, more than any other entity, to package that information in the most persistent form to human memory: the narrative. 
With this in mind, I am a little disheartened that journalists have to resort to sensor journalism to conduct our investigations. After all, should the public, and our sources, need the fourth estate enough that they would just give us the data? But the political economy of the US doesn't work that way, and the media (rightfully so) can't always be trusted. So I am grateful we now have sensor technology at our disposal to help with the heavy lifting.

From listening and taking part in the conversation at Columbia, I wouldn't be surprised if a number of those important sensor journalism initiatives come out of the Tow Center.