Data is the Real Post-Truth, So Here’s the Truth About Post-#Election2016 Propaganda
🏀My Final Post Re: the Post-Fact Era
How can we still be speaking of "facts" when they no longer provide us with a reality that we all agree on? The problem…www.nytimes.com
*It’s done: I’ve added this final round of analysis and network visualizations based on 1.3 million hyperlinks captured across the “fake news” ecosystem and the right-wing’s micro-propaganda machine.*
It feels like someone hit the turbo bass switch and turned up the volume on America’s post-election outrage boombox. Is there a fake news emoji yet?
Pre-facts, post-facts, or no facts, there’s so much misinformed debate, finger-pointing, and political maneuvering right now that I decided a Game 7 was needed to tackle what’s become the great fake news crisis of 2016. And I’m still wondering why more people — notably academics and researchers — haven’t been chipping away more in this election-hacking post-factual super-controversy.
In a New York Times op-ed, William Davie, an associate professor at University of London, elaborates on how we ended up here:
We are in the middle of a transition from a society of facts to a society of data. During this interim, confusion abounds surrounding the exact status of knowledge and numbers in public life, exacerbating the sense that truth itself is being abandoned.
Once numbers are viewed more as indicators of current sentiment, rather than as statements about reality, how are we to achieve any consensus on the nature of social, economic and environmental problems, never mind agree on the solutions? Conspiracy theories prosper under such conditions. And while we will have far greater means of knowing how many people believe those theories, we will have far fewer means of persuading them to abandon them.
Thinking about fake news doesn’t mean extracting the hashtags, the ratios of [😆>😡>❤️] for Facebook posts, or using multivariate statistical analysis to estimate the number of retweets Donald Trump’s transition team updates get each day. All of this might be interesting, but it’s not very relevant to solving the problems. We need serious insights to get past fake news, and this area involves a better understanding of the infrastructure (i.e., the internet) we’ve built that has helped to enable it.
Data is the truth, the post-truth, and nothing but the truth. So help me, data.
Plotting millions of Facebook reactions to a fake news story and hashtag-surfing social media analytics dashboards looking for answers isn’t really helping right now. As a society, we desperately need to start thinking about “insights” that are beyond the gravitational pull of the social networks, the filtering algorithms, and the privacy Death Star.
What Network Maps Show: The Interaction Ecosystem
The linking patterns in large-scale network graphs can be used to analyze the ecosystem of interaction instead of the interaction ecosystem. In the case of fake news, looking closely at these large-scale networks shows us the environments that nurture on-demand news outrage culture and refine its high-octane fuel, misinformation. These types of methods provide us with a much deeper perspective than might otherwise be found using platform-based data and cookie-cutter (pun intended) audience measurement tools.
Everyone’s mad, sad, angry, happy, LOL, WTF, and IDK — I think we all get it. Does it matter if we find out a controversial Trump Facebook post got 100,000😡 or 1,000,000😡? Or who used the #Trumpkin and #WitchHillary hashtags on Halloween? These metrics are loaded anyway— I mean, we are trying get insights from six cartoon emojis meant to capture millions of people’s feelings on complex, often misleading, and frequently rage-inspiring news. Plus, the totals involve network effects: The more followers-friends-bosses-creepers-troll-armies you reach, the more people and bots you can potentially rage troll.
No followers, not many hurt feelings — at least through social media. 20 million followers, and half of them will😡any given post. It’s fairly easy to see how sentiment analysis has worked out recently, right? I mean, DJT could post that he’s “going to Starbucks to get a caramel macchiato #skimmilk” and get at least 150,000😡s along with a few conspiracy theories about the dairy industry’s role in 9/11 and the Great Recession.
The Game 7 of #2016Election Data Finals
Since my last piece on the tracking infrastructure behind the right-wing micro-propaganda machine, I’ve put together a more comprehensive list of “fake news” sites. I’ve used this expanded list, consisting of 30o+identified fake news actors/far-right propaganda sites (I’m sharing this, just like in my previous projects) to construct a more extensive map of the entire right-wing news ecosystem. I feel this is more or less data endgame territory, at least in terms of showing the structure of this right-wing network. Obviously there’s a lot more that can be taken from these websites directly — especially in terms of the content, URL destinations, along with the interfaces, coding, and design of this large group of sites.
What’s shown in the following “big data” network graphs, for all intensive purposes, is the entire right-wing propaganda machine. I can make this claim with a fair amount of confidence, because the data I’ve collected from my expanded list of 306 fake news sites contains a total of 1,338,017o hyperlink connections this time around and involves nearly three times as many fake news websites as my previous related project. I also decided to look at three of the “battleships” (or destroyers, if you’re the type of person who likes to hate the player instead of hating the game) of the right-wing news ops — Breitbart, Infowars, and Fox News.
The beauty of network graphs is that they can be used as structural connections (i.e., links), to find the underlying stories. I don’t care much about what exactly this right-wing fake news and propaganda machine is sharing on its sites every day. Besides, I think I already have a good idea from my previous work. I’m concerned with how these sites are connected and have formed the right-wing news “Voltron” — what’s arguably the most powerful propaganda machine in the history of U.S. election politics.
As I’ve not-so-subtly emphasized in the title of this piece, I feel there are no post-facts in focused and transparent data analysis. I don’t mean to imply that what I’ve done in this project involving fake news is to use data to try and answer every question; it’s just that data is part of the answer to the question at hand. Is there anything reporters could use right now more than a set of graphs that show nearly every known right-wing “fake news” site across an entire cross-section of the internet? Or graphs showing the frighteningly-complex-yet-impressively-coordinated behavioral tracking and social API infrastructure behind it? The same one that’s lying directly underneath the propaganda network? If there is, I’m all ears. So, how about we get back to arguing about politics once we’ve looked at the data?
The Expanded Map of the Micro-Propaganda Machine — Including Breitbart, Infowars, and Fox News
I’ve used two different layers of the propaganda network for this analysis: First, I present this network at the “local” actor level, meaning looking at the links shared exclusively between the group of 306 identified micro-propaganda sites. I turn my focus towards a wide-angle view to paint a picture — quite literally — of what the entire right-wing news ecosystem looks like. The graphs here are based on simple and straightforward techniques, including looking at the In-degree, Out-degree, and Pageranks and force-directed layout approaches. All of this work helps add to the body of evidence I’ve been trying to capture the entire time: to find the sites in this propaganda network that are linked into the network as well as the sites that are the most influential in the connecting.
In total, 1,338,017 hyperlinks were captured in Game 7 of the #fakenews data finals. The content of 21,662 different pages from 306 sites was indexed into using free and open source data tools. Sure, this is big data, but more importantly, it’s reliable data: all the evidence for this discussion comes directly from maps that show the links in, out, and everywhere in between thousands of websites that facilitate the spread of misinformation.
I. The Core Group of 306 Fake News Sites
I’ll work backwards this time, and start with the “close-ups” of the propaganda machine’s core actors (306 websites). Then I’ll zoom out to the larger extended propaganda ecosystem. This ecosystem is the “friends and family” of the original group of 306 sites, which is why the larger big data network graphs in this follow-up project is potentially rich in terms of generating #Election2016 evidence.
The Connectors Linking The Expanded Micro-Propaganda Machine
Looking at the core group of actors — Group 306 — to identify the sites that are front-and-center in the right-wing’s fake news “local” network (see below), there are two easily distinguishable sites linking OUT to other locations in the same zip code. In the graph below, the larger the circle, the more hyperlinks point OUT of the site elsewhere within the same network. Put simply, this really means that these sites are the hubs that potentially direct information-seeking traffic across the thousands of other locations, or nodes, in this micro-propaganda machine.
As you can see, Conservapedia and Rense stand out from most of the others. The Conservapedia graph suggests that this site has become a sort of right-wing Wikipedia for propaganda. I suppose there has to be some kind of wiki to keep of all these new post-facts sorted. In the second graph below, even though I’m not very familiar with Rense, it’s clearly linking OUT to lots of other sites in the core group of propaganda sites. Rense is on the Breitbart and Infowars side of the network, which could be due to the type of content that is being shared (and linked to) on the Rense website.
Destinations >>> Where Micro-Propaganda Sites’ Links Point Towards
The following images show the places in the right wing’s micro-propaganda network where a lot of the links (and naturally, more website traffic) might head in the right-wing propaganda machine. It appears to be nearly identical to the analysis I did for the group of 117 websites: YouTube is clearly the number one right-wing propaganda linking destination. As before, it would be interesting to look at exactly which videos and channels are the most common across this group of 306 websites. Pulling data from YouTube’s API might be the most productive way to engage further with this area. I think the destinations/URLs that these sites are linking to on YouTube are key, because it’s like looking into the propaganda machine’s cable TV package. Except that Comcast doesn’t have an Infowars/ Prisonplanet/ Breitbart add-on. Yet.
II. The Right-Wing News Ecosystem
Now let’s zoom out from the YouTube-linked propaganda solar system of right-wing micro-propaganda and look at the larger news galaxy. Basically, the images below are the right-wing news “internet.” The images below include the 306 websites that are linking to each other as before, as well as the sites that are linked into the larger right-wing news ecosystem. I should point out that not all of these additional sites are fake news. But as the graphs are based on the 306 actors that are connected to them through hyperlinks, it does mean these sites are are connected to one another.
Looking Through the Right-Wing News Ecosystem’s “Crosshairs” (where thousands of hyperlinks are directed)
This larger ecosystem-scale network consists of 767, 382 different web pages. This is the right-wing media ecosystem of propaganda.
Even when zoomed out and despite thousands more sites and links, the results of this graph are fairly similar to the smaller localized network. YouTube is still clearly the destination for links in this larger right-wing news ecosystem. Yet, these views of the network, shown at a scale quite a bit larger than the previous views of the network, can help to provide insight as to the EXTERNAL sites that the smaller group of propaganda sites are connected to (via hundreds of thousands of hyperlinks).
The graph directly below is probably the most fascinating of any I’ve generated in this project: It shows shows the relative placement of the mostly MSM sites that the entire right-wing ecosystem is linking INTO. The red bubbles (i.e., link targets) include the Washington Post, The New York Times, Google, CNN, and Fox News on the lower left. I’d love to look into why (network-wise) Fox News is split off from the cluster of other mainstream media/newspapers that are being linked INTO by the extended right-wing news ecosystem. This however would involve a different type of node weighting related to network centrality.
Same News Ecosystem — Zoomed In
A Close-Up of the Link Targets OUT of the Right-Wing News Ecosystem
The MSM, Trump’s Official Twitter and Election Campaign Website, Breitbart, Fox News, and Wikipedia are shown above in greater detail. Below are a couple of related snapshots in this group that show the “shortest” network paths in an isolated view “TO” Infowars and Wikipedia, a couple of “target” sites of interest of mine that are shown in the previous graphs (see images above). These two bonus graphs are interesting, but mostly meant to be food for thought, since there are lot of other factors at work at this level that help determine what exactly these isolated (“path”) connections mean. Still, these give another perspective on the sites in this “in crowd” of actors at the center of the right-wing news ecosystem.
Network Center — Isolated Path Snapshots
IV. The News Ecosystem Sites That Are Important In Linking Out
Last but not least, in the image below you’ll find the extended right-wing news media ecosystem from the point of view of the sites that play a part in directing links OUT to other websites. On this larger scale, this includes sites in the core group of 306 sites, as well as the links that connect outward to the greater news ecosystem.
It’s fairly similar to the localized OUT network shown earlier. Conservapedia is still tops as a link distribution hub, but in this larger network graph, which contains tens of thousands more link connections, Nodisinfo.com appears to be close to challenging Rense for the #2 position for links going OUT to other places. It’s a bit further from the network’s “center,” though, so probably has a more specific base of visitors. So this ecosystem-scale graph is about the same as the localized version, just with more fringe actors — meaning the conspiracy-focused websites that are distributed around the center of the right wing news galaxy.
To wrap things up data-wise, the website with the most potential outbound reach this network is Conservapedia. While Out-degree informs to some point the connections Conservapedia make across its network of actors, it does not measure how influential the connections are in terms of the relative importance of the sites it is making these connecting to. Despite this fact, 2,908 is a big number, and given Conservapedia is a wiki- informational resource, it’s a strategic hub within the localized micro-propaganda machine, as well as for the larger right-wing news ecosystem. This post is long enough, so I leave it at that.
On the link/connection receiving side of the network was YouTube, which had 182 links coming into the site from the other sites in the list I used for the core group analysis. I’m probably stating the obvious, but it appears YouTube is more important (in relative terms) as a content provider/video archive/cable TV for the micro-propaganda network than it is for the larger right-wing news ecosystem. This, of course, would make sense — a lot of the larger MSM orgs tend to keep their video content in-house.
Re: Keeping It Real
In conclusion, my last several Medium posts have done a lot of data work on the networks that underpin “fake news.” I’m not trying to pick sides here: the entire point of this project was to see what a large and highly relevant sample of “big data” could tell us about fake news. I’ll admit, there is focus on a certain group of actors: the right-wing news media and the smaller, more misinformation/propaganda-focused group of websites. But my intention is for this data to shed light on future conversations about “fake news,” not to detract from them. Having said this, I’m taking a break until next season. This was my Game 7 of the data finals. I’ve spent many hours on this series of #fakenews posts, and have tried to openly share my sources, lists, and research approach for a reason: so that this work might contribute to evidence through a straightforward and transparent data analysis to what seems to me is a widening controversy.
Reporting on “fake news” with unsubstantiated claims and incomplete evidence isn’t the best approach — in fact, it’s probably the worst strategy, because it adds to the existing noise. Lately, I feel that many journalists and news organizations are churning out news in response to the public, rather than leading the way to inform them on the underlying issues. In the post-fact era, what’s in the public interest seems to be the opposite of what the public is interested in. Right now—and for the foreseeable future, good data, just like Paul Pierce during NBA playoff time — is the real post-truth. News media need to move away from selling their products based on people’s attention and staring finding ways to sell them The Truth.