FashMaps: Locating White Supremacists in Space

Are you worried that your neighbors are actually Nazis? Do yowling white men in polo shirts with suspect haircuts continue to hide from you, no matter how carefully you scrutinize your neighborhood? Seek no further.

There is now “Fashmaps,” an activist-run website that uses public web postings to figure out where white supremacists claim to be located and where they will be congregating for meet-ups. Each point on the “Nazis in your Neighborhood” map (imprecisely) locates a user of the infamous Daily Stormer Neo-Nazi web forums. It’s possible because some Daily Stormer users knowingly post their locations, while some of their meet-ups have publicly viewable locations.

The map displays only locations and meet-up events that are freely and public posted on the Daily Stormer website. It pointedly excludes additional, corroborating information that might be used to locate individual users more precisely: the website clearly states that it is not intended for violence, stalking or harassment. All it purports to do is to locate American white supremacists in place and time, to document the spatial realities of our current, burgeoning alt-right problem.

It may surprise you that Daily Stormer users are so willing to publicly post their locations. It shouldn’t be a surprise. Our modern-day fascists are surprisingly poor at — or so confident as to be uninterested in—operational security.

Consider how the delightfully-named Unicorn Riot media collective was able to obtain (via an anonymous source) and publish logs from the Discord chatrooms used by the Charlottesville tiki-torch wielders. These logs are now being used by lawyers in the court-case against the Charlottesville planners, strengthening the argument that their actions constituted a criminal conspiracy. Just like Daily Stormer users, the Charlottesville planners failed to even consider that someone might be watching them.

I’ve read many dim-witted and oh-so-earnest tactical conversations in advance of anti-Trump protests and dubiously-named “free speech” rallies on public 8Chan boards: they are often as self-aware as that spider from the video industriously covering itself in sand, blissfully ignorant of the fact that we can still see it.

In these conversations, there is always lots and lots of back-and-forth about making Secret Plans that AntiFa and the Liberal Media Will Never See Through: I’d scan through these postings and I’d wonder: “Aren’t they aware that I can see this, on this public and widely-known board? Don’t they know these boards aren’t an exotic secret? Do they lack a theory of mind or something?” (Well, yes, to that last one)

Perhaps it’s because they’re newly emboldened, and feel they no longer have to cover their tracks. Perhaps it’s because the FBI — as was recently demonstrated by their total failure to act on a tip about the school shooter in Parkland — are still bizarrely incapable of following up on potential criminal activity if people are talking about it in the great and mysterious land of Online. Perhaps it’s because many people involved with these online-hate groups simply haven’t realized that doxxers can also be doxxed in return.

Is FashMaps effective? Well, that depends on what we mean by “effective.” Do we want the tool to discourage white supremacists by exposing their (general) locations, or do we want it to be effective as a tool for awareness and research tool? Insofar as I can tell — and I’m going to do more digging—it’s decidedly unclear if the threat of being mapped helps to actually discourage and disorganize white supremacists. While the Southern Poverty Law Center regularly is complained about (and occasionally sued) by the groups it names in its Hate Map, it’s not clear if the Hate Map actually deters white supremacist activity — such a causal link might be a real pain to prove, anyway.

We don’t know if the threat of being located in space is a deterrent to white supremacists. We do have decent anecdotal evidence that the threat of being doxxed —in which someone’s personal information, such as their address, is dug up and widely disseminated online—has “terrified” many white supremacists. It will be extremely interesting to see if fewer white supremacists and alt-righters turn up to real-world marches and rallies in 2018, as left-wing activists step up their efforts to publicly expose them.

We have more concrete evidence that mapping projects like FashMap are potentially useful tools for awareness and research, giving law enforcement, policy makers, and researchers a better way of visualizing and drawing connections between hate groups.

The Southern Poverty Law Center’s HateMap.

For starters, it’s not exactly the first such effort to map hate and violence online: it’s part of a very long tradition of web-mapping for some kind of political or activist purpose. The Usahidi web-mapping platform was launched in 2008 to track political violence during the Kenyan elections, and is still being used today: it served as the inspiration for the roughly-gazillions of similar interactive mapping projects that popped up after it. The still-running “Rechtes Land” web map, created by a data journalist, has tracked Neo-Nazi activity since 2017.

Visualizing the spread of hate is powerful, and maps represent a potent and easily-understood way of performing that visualization. Hate groups are “geographical phenomena,” as this 2017 paper from the Annals of the Association of American Geographers states: it goes on to observe that “The connection of hate and place stems from the social construction of place and its link with identity formation and stability (Gallaher 1997).”

The aforementioned Southern Poverty Law Center HateMap may or may not discourage white supremacists, but it’s indisputable that it’s one of the best existing resources for reporters and researchers who want to better understand American hate groups. FashMaps is really just another interesting instance of how cartography and the aggregation effect can be combined, can even be turned into a sort of political weapon.

Is FashMaps ethical? Yes, I think it pretty much is. As criminology professor Brian Levin acknowledged in a Vice article about FashMaps, the DailyStormer’s Internet Nazis are revealing their identities in the “public square” of the Internet. It’s the equivalent of posting the location of your gross Nazi party in a mimeographed flier stuck to a telephone pole: if you don’t want to be located by people who don’t like Nazis, you probably shouldn’t be willy-nilly sharing your location. There is no reason why Neo-Nazis, who (lest we forget, which a lot of us weirdly seem to these days) vocally wish to exterminate their fellow human beings, should be permitted more geographic privacy than anyone else. The law seems to agree with this as well, at least for now. Legal challenges to the SPLC’s Hate Map from the named groups, for example, have largely hinged on defamation and trademark violation, not on the actual act of geographically locating these groups.

What’s more, FashMaps is locating these DailyStormer users in an extremely gentle and considerate fashion. FashMaps could be much less gentle. That’s because it’s usually very easy, trivially easy, to find out exactly where most Internet users live, work, and meet up. Most people — including Neo-Nazis — just don’t understand that security online is more difficult than setting up two-factor authentication and not using your dog’s name as a password. They don’t realize that it’s possible to find somebody, find out a lot about somebody, by just linking a few different sources of publicly-available, freely-given and seemingly innocuous information.

I just wanted an excuse to add a funny Roman mosaic picture.

This is sometimes called the “mosaic effect”: it’s what you do when you combine multiple sources of data to reveal a full picture about a individual. In today’s social media world, where people willingly share scads of information about themselves and where they’re going and what they’re eating, pulling together these bits of information requires only a bit of deductive ability, not technical brilliance.

FastMaps and the ethical questions it evokes can be plugged comfortably into the much larger, long-running debate surrounding the ethics of aggregating publicly-available data in ways that the initial posters may not have anticipated. It’s actually almost quaint, considering that it’s not in any way automated, does not involve an algorithm or artificial intelligence or any of the stuff that today’s ethicists are currently verklumpt over: per the site owner, it’s just one guy and some volunteers combing through Daily Stormer for location clues.
“Fine, whatever, but won’t the Neo-Nazis and their sympathizers turn around and geolocate us?” you ask me.

Good point: but unfortunately, that’s already happening. In 2017, online alt-righters used a public petition from anti-Trump group RefuseFascism to collect and disseminate the addresses and personal information of hundreds of signees. A German Neo-Nazi group released a “Jews among us” web map on Facebook in 2016, which mapped “synagogues, day care centers, schools, memorials, businesses, restaurants and cemeteries.” The American Family Association created a “Bigotry Map” in 2015 that located LGBT organizations, though it vanished in just a few months.

FashMap isn’t particularly novel and we really have no way of knowing if it will actually deter the Daily Stormer’s user-base from doing horrible things. I am still glad it exists. It has prompted me to think about how the Internet has facilitated these map-against-map conflicts, or made them much more visible. What happens when we’re all busily attempting to out-locate the other? And why are we still — everyone, not just Neo-Nazis — still so incredibly bad at defending ourselves against being mapped when we don’t want to be?

Consider the recent outrage over the Strava activity-mapping service, which even people who ought to be highly security minded — soldiers on patrol, foreign agents — inadvertently permitted to track their activities. While we recognize the power of maps, as those howling over the SPLC Hate Map obviously do, we still willingly share plenty of geographic information about ourselves.

We are all eagerly mapping each other, and we are all still so terrible at not being seen.

Drone Mapping a Mental Hospital with the DJI Phantom 3 Professional

Medfield Mental Hospital from the air.
Medfield Mental Hospital from the air.

I recently bought a Phantom 3 Professional, operating under the logic that it costs $1200 and is therefore much more economical than a hexacopter. Myself and my partner, Daniel, are working on developing expertise in 3D mapping with a UAV, and I’d been looking for a new model capable of waypoint navigation and shooting high-quality, undistorted still images. My Phantom 2 still worked great, but it wasn’t great for mapping – built to use a fish-eye lens GoPro camera, and unable to carry out waypoint navigation without extra, expensive parts.

I was really sold on buying a Phantom 3 Pro after I visited the DroneDeploy offices in San Francisco and watched a demo of their waypoint navigation software, which is paired with their cloud computing processing. You fire up your mobile phone or tablet, sync it with the Phantom 3, then draw a box around the area you want to map. The software calculates how many times the Phantom will need to cross the area, the altitude of the area, and how many pictures are required, then you press a button. The Phantom proceeds to launch itself and carry out its work without your input, though you can always call it back from the controller. Simplicity. I like it.

So, I bought a Phantom 3 Pro—  and since I live in the giant no-fly-zone otherwise known as Washington DC, I had it shipped to Daniel in Boston where I regularly visit him. On my last visit in early September, we decided to test out DroneDeploy and the Phantom 3 by using it to map the abandoned Medfield State Hospital  in Medfield, Massachusetts, which I’d found out about on Atlas Obscura. (Scenes from “Shutter Island” were filmed there). Unlike most creepy, abandoned mental hospitals, this one had been opened to the community for use as a park, while the town decides how best to redevelop it. It’s a sprawling complex with red brick architecture and lush greenery around it in summer, with the Charles River bending towards one corner.

My new Phantom 3, configured to run DroneDeploy off my Galaxy Note 8.0 tablet.
My new Phantom 3, configured to run DroneDeploy off my Galaxy Note 8.0 tablet.

We parked across the street and walked in, and identified a parking lot where we could easily launch the drone from a flat location. DroneDeploy synced up easily enough with my Phantom 3, and I chose to map about half of the area, going conservative for a fist-time experiment. I pressed the button. It worked great: the Phantom efficiently flew off in the designated pattern, in  neater lines then I could manage myself.  It retuned to home in about 15 minutes, and landed itself, albeit with more force then I’d like. I may, in the future, switch back on manual control of the Phantom as it comes in to land after a DroneDeploy mission, as I prefer to catch it rather than landing it.

Since DroneDeploy missions currently can’t be flown with the camera at an oblique angle, I manually shot my own oblique imagery, with the Phantom 3 camera set to shoot images every five seconds. I flew reverse transects from the DroneDeploy pattern, and – following advice from DJI’s Eric Cheng – flew the drone in large, slow circles over the area I want to map. I probably should have worked with alternating the altitude more, but I was pleased enough with the images I was able to collect. The Phantom 3 handles even more smoothly than the Phantom 2, and shoots beautifully crisp still images with its 12-megapixel camera, without the distortion that used to annoy me with the GoPro.

We used both DroneDeploy’s processing tool and Agisoft Photoscan 3D to process the final imagery. Daniel has a great summary of the pros and cons of each over at his blog, so I won’t recap them – but in summary, DroneDeploy was a lot faster, while Agisoft PhotoScan had higher quality results but took a longer time and required much more processing power, and also required us to manually fill in some holes in the mesh.

Here is the final, orthorectified map. DroneDeploy’s ability to quickly orthorectify 2D maps using cloud processing is definitely handy. In the 3D model, DroneDeploy was not able to incorporate our oblique imagery successfully, although we’ve been in touch about the problem, and they’ve told us it will be fixed. There’s two other problems with DroneDeploy as of this writing: it only works with Android phones and tablets, and it requires either Wifi access or mobile data to function.

Both features are in the works, but keep this in mind if you want to experiment with it.In Agisoft Photoscan, which did use our oblique imagery, the sides of the model weren’t as detailed as we’d like – though, some of this is to be expected when mapping an entire complex of buildings.  We could probably fix this by taking the time to shoot oblique imagery around each individual building, but this would take quite a bit of extra time and battery power. (I’d like to try it anyway).

The Drone Deploy model:

Medfield State Hospital
by mountainherder
on Sketchfab

The Agisoft Photoscan model:

Medfield State Hospital – PhotoScan
by mountainherder
on Sketchfab

Overall, I’m very pleased with the Phantom 3 Professional as an inexpensive mapping tool, and I’m excited to see what we can come up with next.  I’m also interested in doing more work with DroneDeploy – and I eagerly await the release of the off-line version, which should make it a much more viable tool for field work. What else could we map in the area around Boston?

Drones and Aerial Observation: our primer for New America is finished!

drones drones drones

 

We’ve finally done it: the “Drones and Aerial Observation” primer I’ve been working on for New America with support from the Omidyar Network and Humanity United has been released into the wild.  Ever wondered how drones can help with peaceful endeavors, from disaster response, to conservation, to archaeology? We have you covered.

With this book,  I’m of the mind that myself, my colleague Konstantin Kakaes, and the drone experts who contributed chapters have created an overview of drone technology accessible to people who don’t already know what a “gimbal” is. (Yes, I am aware that is a funny word).

We hope the book will encourage people to start thinking of drones as a tech they can practically use for their own field endeavors. While drones certainly look complicated when you first encounter them – at least, that’s how I felt about them – it’s a tech that’s remarkably accessible to people who don’t have aeronautical engineering PHDs.

You can download the whole shebang as a PDF,  or you can also download individual chapters. Share it, print it out, tell your friends, tell us what you think, tell your friends what you think.

On my end, I wrote chapters 4 and 5: “How to Make Maps with Drones” and “Mapping in Practice.”  Writing these chapters was a real crash-course in drone mapping for me, and I’m grateful to come out the other side alive and with a better sense of what’s required to carry out mapping projects. I hope I can pass that on to you. I’m also planning to get my own mapping drone in the very near future so I can start carrying out some of this work myself.

I also wrote Chapter 9, which is a case study of the world’s largest archaeological drone mapping project, carried out by the Ministry of Culture in Peru. They were incredibly hospitable to me,  and I had a great time watching the researchers deal with the quotidian, difficult, occasionally terrifying realities of making maps with drones in remote and difficult areas. Many thanks to Aldo Watanave and Dr. Luis Jaime Castillo Butters for taking me along for the ride. A Slate piece about this work is impending as well.

To celebrate the release of the book on July 22nd, we held a “Drones and Aerial Observation” symposium at our Washington DC offices. The half-day event featured a lot of great thinkers and practitioners on UAV technology, and from my admittedly biased perspective, I thought it went very well. You can see videos and slideshows of the panel discussions at this link. 

I’d love to hear what you think about the primer, so feel free to reach out to me on Twitter or Facebook, or maybe even email. More drone-related writing and research coming up: watch this space!

dji S1000 pisaq BW
My favorite photo from my distinctly drone-focused trip to Peru.