<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by a data scientist on Medium]]></title>
        <description><![CDATA[Stories by a data scientist on Medium]]></description>
        <link>https://medium.com/@anonymousdata?source=rss-9aad9a2ae2dd------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*IEHAP4SmSALXJUNhloKlRg.jpeg</url>
            <title>Stories by a data scientist on Medium</title>
            <link>https://medium.com/@anonymousdata?source=rss-9aad9a2ae2dd------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 03:38:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@anonymousdata/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Postman is logging all your secrets and environment variables]]></title>
            <link>https://anonymousdata.medium.com/postman-is-logging-all-your-secrets-and-environment-variables-9c316e92d424?source=rss-9aad9a2ae2dd------2</link>
            <guid isPermaLink="false">https://medium.com/p/9c316e92d424</guid>
            <dc:creator><![CDATA[a data scientist]]></dc:creator>
            <pubDate>Fri, 16 May 2025 09:34:32 GMT</pubDate>
            <atom:updated>2025-05-16T16:40:31.147Z</atom:updated>
            <content:encoded><![CDATA[<p>I was originally investigating <a href="https://www.stedi.com/blog/postman-is-probably-not-hipaa-compliant">this report</a> that Postman is not HIPAA compliant. I found that Postman is not just wholly unsuitable for anyone testing a healthcare application — it has virtually zero regard for the privacy of <em>any</em> of its users, and has probably logged every secret string you have ever given it.</p><h3>Charles Proxy</h3><p>This investigation would not have been possible (or so effortless) without <a href="https://www.charlesproxy.com/">Charles Proxy</a>. It is $50 for a lifetime license — when you cancel your $49/month enterprise seat at Postman after reading this post, please consider supporting truly ethical software developers by purchasing a license, even if it’s just to verify my findings for yourself.</p><p>The first thing you’ll notice is that Postman logs <em>a lot of stuff</em>. Before you even interact with the app in any way, hundreds of network requests have been fired off to a constellation of analytics tools and third-party integrations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qjksUvdaHsao84daesqblw.png" /></figure><p>This is actually quite common in modern software, which tends to follow the ethos of <em>log everything</em>.</p><h3>Certificate pinning</h3><p>To inspect the encrypted traffic flowing to Postman endpoints, you will need to enable <a href="https://www.charlesproxy.com/documentation/proxying/ssl-proxying/">SSL proxying</a> and bypass the <a href="https://security.stackexchange.com/questions/29988/what-is-certificate-pinning">certificate pinning</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iILOehRFN-vNQm4UAMcyyw.png" /></figure><p>Once you set up a spoofed root certificate, you can start to see the decrypted payloads that are being sent to Postman’s servers.</p><h3>Postman leaks secrets</h3><p>This was so trivially easy to find that I was genuinely surprised nobody else is making a big deal about this. If I create an environment variable and set it to “secret”, it might be hidden from the rest of my team, but it’s definitely not hidden in the logs sent to Postman.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SUib27vyr2laP-HK4PdheA.gif" /></figure><p>It doesn’t matter if you set an environment variable as secret or not — whether or not it is masked on the screen, Postman will capture it via the resolvedRequestUrl string, which captures the state of the URL string <em>after inserting secret variables</em>.</p><p>The Postman UI has a popup with a bold claim, that variables are a way to “protect sensitive data” — this is a verifiable lie if my supposedly sensitive data is being sent to your logs in plain text.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wfNdKEYCGv7OT9G3Nai2iQ.png" /></figure><p>Just to be precise on what the problem is here:</p><ul><li>If I create a secret string, I don’t want you to send it to your company servers. I don’t see how that is an unreasonable expectation.</li><li>While it seems like some care was taken to not log any headers, it is entirely possible for PHI and sensitive data to end up in Postman logs from the URL and query parameter logging.</li></ul><h3>If you really must use Postman</h3><p>At least add this to your /etc/hosts file, so the analytics endpoints that are logging your secrets stop receiving them.</p><pre>127.0.0.1 bifrost-https-v4.gw.postman.com<br>127.0.0.1 bifrost-v4-global.gw.postman.com</pre><h3>A note about ethics</h3><p>Why is there no Hippocratic Oath for software developers? Postman is probably not the only tool where the employees are stashing your unencrypted strings in unencrypted places.</p><blockquote>Now if I carry out this oath, and break it not, may I gain for ever reputation among all men for my life and for my art; but if I break it and forswear myself, may the opposite befall me.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9c316e92d424" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The CDC reported 11.3 million wasted COVID-19 vaccine doses in 8 months]]></title>
            <link>https://anonymousdata.medium.com/the-cdc-reported-11-3-million-wasted-covid-19-vaccine-doses-in-8-months-5bdff71a850a?source=rss-9aad9a2ae2dd------2</link>
            <guid isPermaLink="false">https://medium.com/p/5bdff71a850a</guid>
            <category><![CDATA[cdc]]></category>
            <category><![CDATA[vaccines]]></category>
            <category><![CDATA[vaccine-hesitancy]]></category>
            <category><![CDATA[waste]]></category>
            <category><![CDATA[covid19]]></category>
            <dc:creator><![CDATA[a data scientist]]></dc:creator>
            <pubDate>Sun, 05 Sep 2021 08:10:16 GMT</pubDate>
            <atom:updated>2021-09-05T08:10:16.939Z</atom:updated>
            <content:encoded><![CDATA[<p>By the end of July, the U.S. Center for Disease Control (CDC) had administered nearly 368 million doses of the COVID-19 vaccine. In the process, at least 11.3 million doses (~3%) were wasted.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YetEMJUVdAiljddKEJ5AxQ.png" /><figcaption>Total number of wasted doses by week, which has been steadily growing since April.</figcaption></figure><p>The CDC data is missing entire states like Michigan and Connecticut, federal vaccination programs like the Department of Defense and U.S. Bureau of Prisons, and various smaller vaccination programs that do not report wastage to the CDC. As of August, <a href="https://www.nbcnews.com/news/us-news/america-has-wasted-least-15-million-covid-vaccine-doses-march-n1278211">at least 15 million doses have been wasted</a>, with the full scale of wastage yet to be accounted for.</p><h3>How vaccines get wasted</h3><p>The mRNA vaccines developed by Pfizer, Moderna, and J&amp;J must be kept at a specific low temperature during transport and storage, so any refrigeration problems can cause a mass wastage of doses. The Moderna and Pfizer vaccine must be administered within 6 hours of removal from temperature control, while the J&amp;J vaccine must be used within 12 hours. Some vaccines come in vials of five doses each, while others come in vials of ten, leading to waste when a full vial cannot be used in time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*svznG0R-zXaaz5L31n69eA.png" /><figcaption>Wastage grouped by vaccine manufacturer. 56% of wasted doses were of the Moderna vaccine, 32% were Pfizer, and 12% were from J&amp;J.</figcaption></figure><p>The CDC released guidelines on how <a href="https://www.cdc.gov/vaccines/covid-19/downloads/wastage-operational-summary.pdf">vaccine providers should minimize waste</a>, and the WHO recommends keeping waste below 1%.</p><h3>Where vaccines get wasted</h3><p>More than half of all waste occurs at Walgreens, CVS, Walmart, Rite Aid, and major retail pharmacies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/584/1*GEoKTVtzRlGlP3FGd3WHLg.png" /><figcaption>Waste by vaccine provider. Large retail pharmacies account for around two-thirds of all <strong>reported </strong>vaccine waste — state vaccination programs, hospitals, clinics, and smaller providers account for the remaining third.</figcaption></figure><p>When an instance of waste is reported to the CDC, the provider name includes the pharmacy store number (e.g. “WALGREENS #123” or “RITE AID 00120”). Store locators like the <a href="https://www.walgreens.com/storelocator/find.jsp">one for Walgreens</a> make it easy to programmatically retrieve the latitude/longitude of a particular store, which can in turn be used to generate a relatively high-resolution heat map of vaccine waste in the United States over time. I’m working on mapping all retail pharmacy locations to waste over time — if you have data on pharmacy locations (ideally with latitude/longitude), I would greatly appreciate if you could <a href="/service/mailto:contact@adatascienti.st">email them to me</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aCNxALrDknJasvU0dCrr5Q.png" /><figcaption>I am currently exploring the data with the Wolfram Language, which is how I found that the “submitted” date is not exactly correlated with the actual day a dose was wasted.</figcaption></figure><h3>FOIA request</h3><p>Data on vaccine waste was not publicly disseminated anywhere, so I <a href="https://foia.cdc.gov/app/Home.aspx">filed a FOIA request with the CDC</a> for any relevant information.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/690/1*BuMrtsn64EibYAfa83V90A.png" /><figcaption>I did not qualify for a fee waiver since <a href="https://adatascienti.st/">my website</a> seemingly wasn’t enough to demonstrate that I disseminate information to the public.</figcaption></figure><p>My request finally received a response containing a massive Microsoft Excel 2007 spreadsheet with a row for every reported instance of waste. You can avoid the horrendous CDC interface for FOIA requests, and the difficulty/cost of this process, by downloading the <a href="https://adatascienti.st/cdc/wastage.xlsx">raw data I received from the CDC</a> as well a <a href="https://adatascienti.st/cdc/wastage.ndjson.zip">nicely-formatted NDJSON file</a> where I made a <em>first attempt</em> at parsing the values and cleaning things up.</p><p>You can also use the aws command line utility to download this dataset and all future updates.</p><pre>aws s3 cp s3://adatascientist/cdc/wastage.ndjson.zip .</pre><h3>Support data science</h3><p>You can support this project and ongoing work like it by</p><ul><li>visiting <a href="https://support.adatascienti.st/"><strong>https://support.adatascienti.st</strong></a></li><li>emailing me your thoughts at <a href="/service/mailto:contact@adatascienti.st">contact@adatascienti.st</a></li><li>contributing your time in analyzing the data and thinking of other datasets that this data can be referenced to. Is there a link between <a href="https://data.cdc.gov/stories/s/cnd2-a6zw">vaccine hesitancy</a> and vaccine waste, or to some other demographics of a community?</li></ul><p>I’ve had some months since my last post, and have some projects like this one in the works that will get my motivation back to where it used to be. Stay tuned!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5bdff71a850a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The sources of misinformation that thrived in Parler]]></title>
            <link>https://anonymousdata.medium.com/the-sources-of-misinformation-that-thrived-in-parler-cb740a98bd0d?source=rss-9aad9a2ae2dd------2</link>
            <guid isPermaLink="false">https://medium.com/p/cb740a98bd0d</guid>
            <category><![CDATA[misinformation]]></category>
            <category><![CDATA[parler]]></category>
            <category><![CDATA[coronavirus]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[deplatforming]]></category>
            <dc:creator><![CDATA[a data scientist]]></dc:creator>
            <pubDate>Sat, 16 Jan 2021 01:05:10 GMT</pubDate>
            <atom:updated>2021-03-11T20:36:11.362Z</atom:updated>
            <content:encoded><![CDATA[<p>It seems Americans are not on the same page about COVID-19. <a href="https://www.pewresearch.org/fact-tank/2020/07/24/a-look-at-the-americans-who-believe-there-is-some-truth-to-the-conspiracy-theory-that-covid-19-was-planned/">25% of Americans still believe that the pandemic was planned and executed by powerful people</a>. In this post I’ll look at how misinformation spread on Parler and correlate external domains to the conversation happening within the network.</p><blockquote>This is part 3 of a <a href="https://anonymousdata.medium.com/">multi-part series</a> where I analyze data scraped from Parler before it was shut down by Amazon. In part 1, I analyzed <a href="https://medium.com/swlh/violent-hashtag-frequencies-in-parler-eddab2871b66">the frequency of violent hashtags</a>, and in part 2 I got distracted by the <a href="https://anonymousdata.medium.com/the-political-campaign-that-put-a-quarter-million-dollars-into-parler-after-the-election-ee6c526c1d44">bizarre campaign finance records of Marjorie Taylor Greene</a> and the quarter million in dark money that made its way into Parler over the course of a single month.</blockquote><p><strong>As of March 10th, </strong><a href="https://parler.adatascienti.st/"><strong>anyone can search the full dataset</strong></a><strong> and download my full dataset, with all archived pictures and videos from Parler’s CDN. Thank you to all the incredible people on the Internet who helped make this possible.</strong></p><h3>What is misinformation?</h3><p>I occasionally browsed Parler while crawling it, and was fascinated by its “billionaire conspiracies”. In such a conspiracy, a billionaire is involved in some reprehensible secret activity, such as <a href="https://www.facebook.com/57647760528/posts/10163726336570529/?vh=e">this video about Bill Gates</a> shared over 72,000 times on Facebook. The spurious argument is that a multi-billionaire could apply his knowledge of computer viruses and their antivirus software toward creating human viruses and the corresponding vaccine. The comment section is a war zone between those who subscribe to the theory and those who find it ridiculous.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dDgz0uVH1bBFovnzDQ_i_w.png" /><figcaption>Most Parler users are highly critical of Bill Gates, but I didn’t have the means to classify all posts into positive/negative sentiments. It is possible (although extremely unlikely) that there is some Parley out there where a user is simply sharing the humanitarian achievements of the Gates Foundation.</figcaption></figure><p>The <a href="https://www.facebook.com/57647760528/posts/10163726336570529/?vh=e">video I linked to above</a> was released in early April (the left-most red line above), and was barely noticed on Parler. Toward the end of June, a now-deleted Facebook video claimed Bill Gates was planning to use vaccines to depopulate the Earth (second red line). As the national discourse shifted toward vaccine development, a series of <a href="https://www.zerohedge.com/health/robert-f-kennedy-jr-exposes-bill-gates-vaccine-agenda-scathing-report">blatant lies</a> (<a href="https://www.indiatoday.in/fact-check/story/conspiracy-theory-bill-gates-backed-polio-vaccine-disabled-kids-1707666-2020-08-04">debunked here</a>) were widely shared on Facebook and Parler, claiming the Gates Foundation paralyzed kids while distributing vaccines in third-world countries. By December, enough people had died of coronavirus in America that all but the most conspiratorial minds were forced to accept the fact that the disease is indeed out there. Just like the virus, the misinformation around the virus mutated into a <a href="https://www.facebook.com/theresurgent/videos/418364602851481">rallying cry against public safety measures proposed by Bill Gates</a> and a debate as to whether the benefits of opening the economy outweigh the deaths it would cause.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*bEO6rD8dh7UzcWnIYQrtxA.gif" /><figcaption>The monthly word cloud of hashtags on posts that reference Bill Gates in any way.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MU6n4wXX5TRkLitqVzFuvA.png" /><figcaption>A couple lines of Wolfram Language code generated the GIF above.</figcaption></figure><h3>Misinformation is constantly adapting</h3><p>This was a popular post on Parler from September 26th, 2020. As I mentioned in part 1, I will not be publishing any usernames or helping you find who wrote this.</p><blockquote>#OPENAMERICANOW !!<br>We have the Medicines !!<br>We have the logistics/hospitals !!<br>#KidsAreSafe<br>#NoVaccineForMe !!<br>We already are at #herdimmunity <br>#NoMask =Healthy<br>#Mask =Sick, vulnerable or terrorist.<br>#NoContactTracingForMe<br>#Science <br>#billgatesbioterrorist <br>Ask #DrJudyAMikovits and about #FauciTheFraud !<br>#USA USA USA!<br>#ChicanosForTrump<br>#MAGA #MAGA2020<br>Dr. Atlas Breaks Down the REAL COVID-19 Numbers on The Ingraham Angle — Rebukes CDC Director Dr. Redfield on Scare Tactics (VIDEO)</blockquote><p>While many of us got our information from Dr. Anthony Fauci, the rest of America was listening to Dr. Scott Atlas <a href="https://twitter.com/IngrahamAngle/status/1335240859655299072">voice opinions on Fox News that masks and social distancing don’t work</a>. These videos were viewed and reposted thousands of times on Parler.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jXTMBgYCU-ok1OkCuMAVMg.png" /><figcaption>Dr. Scott Atlas’ appearances on Fox News are marked by red vertical lines. I had to search the Fox News website to collect these dates, so there may have been other appearances that were not labeled. The green line on the right is the day he resigned from the White House Coronavirus Task Force.</figcaption></figure><p>Even after Dr. Scott Atlas resigned from the coronavirus task force on November 30th following a Tweet encouraging Michigan residents to “rise up” against newly announced COVID restrictions, the momentum of doubt cast upon lockdowns and masks was enough to keep growing without him.</p><blockquote>I don’t want to say anything against Dr. Atlas as a person but I totally disagree with the stand he takes. I just do, period.<br> — Dr. Anthony Fauci, November 19th, 2020, NBC News</blockquote><h3>Misinformation videos that inspire violence</h3><p>From anecdotal browsing experience, I would see a good amount of outrage in the comments associated with posts that had a video. I first filtered my dataset for posts linking to videos that contained violent hashtags, or whose average comment had a very negative sentiment when passed into <a href="https://aws.amazon.com/comprehend/">AWS Comprehend</a>. I have not evaluated the Comprehend piece thoroughly, but adding/removing it does not significantly change the results.</p><p>I weighted each hashtag by the impressions and comments on its associated posts, then picked the top 100 hashtags and the web domains they led to.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dhGfz4HU_Wk-Ns3jkWYLXw.png" /><figcaption>I wrote a function named ViolentQ (honoring the naming scheme of the Wolfram Language) which first checks for violent hashtag patterns, then passes the post body into AWS Comprehend if a match is not found.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NKJ0Xr48XMu2UqShQXOI0Q.png" /><figcaption>Videos that provoke angry/hateful responses are primarily hosted by these two companies.</figcaption></figure><p>Misinformation videos that evoke a violent response (“hang the traitors”, “kill all of them”, etc) is primarily hosted by two sites — <a href="https://www.bitchute.com/">BitChute</a> and <a href="https://www.youtube.com/">Youtube</a>. Keep in mind that Google, which owns Youtube, was the first to ban Parler from the Google Play store for inciting violence.</p><p>Only after doubling the threshold for hashtag frequency does another video site (<a href="https://rumble.com/">Rumble</a>) begin to appear.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5vUoX7J28jXkc9K1BWocmA.png" /><figcaption>Due to the amount of overlap between video topics on BitChute and YouTube, there is no easy way to separate them in this rendering. However, it becomes clear that videos on QAnon, Sidney Powell’s efforts to “release the kraken”, and the far-right militia group Proud Boys, are predominately shared on Rumble.</figcaption></figure><p>A significant part of America does not know about these sites. Facebook is not suggesting them, Twitter is not recommending to follow them, and Google is not showing them in the first page of results.</p><h3>The Pirate Bay</h3><p>The popular file-sharing website ThePirateBay allows anyone on the Internet to download the plans for a 3D-printable single-shot handgun. Despite <a href="https://en.wikipedia.org/wiki/Liberator_(gun)">significant controversy and US government intervention</a>, TPB defended their decision to host the information.</p><blockquote>We don’t condone gun violence. We believe that the world needs less guns, not more of them. We believe however that these prints will stay on the internets regardless of blocks and censorship, since that’s how the internets works. If there’s a lunatic out there who wants to print guns to kill people, he or she will do it. With or without TPB. Better to have these prints out in the open internets (TPB) and up for peer review (the comment threads), than semi hidden in the darker parts of the internet.</blockquote><p>As long as there is someone out there willing to blame Bill Gates for the pandemic or call George Soros a Nazi, there will be a place on the Internet that is willing to give a voice to their ideas. When these places shut down, the people on these platforms scatter to the next place on the Internet that will take them.</p><p>One question we should be asking is whether we should banish certain ideas to an echo chamber that others have no incentive to enter, or should we allow free expression of all ideas — no matter how repulsive — so some higher ideal for Internet discourse may have the opportunity to prevail.</p><p>You can <a href="https://twitter.com/anonymousdata_">follow me on Twitter</a> where I exclusively post about data science topics. In part 4, I’ll use the complete scrape of Parler from <a href="http://archiveteam.org/index.php?title=Parler">the Internet Archive</a> to follow its users to new corners of the Internet where they probably used the same username. Not going to be naming any names, just looking at the statistics and estimating based on that.</p><p>If you are a researcher or just a curious person, you can email <a href="/service/mailto:contact@adatascienti.st">contact@adatascienti.st</a> to request the data, which I would be more than happy to send to you. I have emailed many people just this week, and still think we need more eyes on this problem. The full scrape of Parler is 1/6th ready if you would like that as well.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cb740a98bd0d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The dark political campaign money that was funneled through Parler]]></title>
            <link>https://anonymousdata.medium.com/the-political-campaign-that-put-a-quarter-million-dollars-into-parler-after-the-election-ee6c526c1d44?source=rss-9aad9a2ae2dd------2</link>
            <guid isPermaLink="false">https://medium.com/p/ee6c526c1d44</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[super-pacs]]></category>
            <category><![CDATA[parler]]></category>
            <category><![CDATA[campaign-finance]]></category>
            <category><![CDATA[politics-and-elections]]></category>
            <dc:creator><![CDATA[a data scientist]]></dc:creator>
            <pubDate>Thu, 14 Jan 2021 09:06:10 GMT</pubDate>
            <atom:updated>2021-03-11T20:36:54.705Z</atom:updated>
            <content:encoded><![CDATA[<p>This is the second part of a multi-part series on data I crawled from Parler’s API until it was shut down by Amazon. I got a surprising amount of attention from <a href="https://anonymousdata.medium.com/violent-hashtag-frequencies-in-parler-eddab2871b66">part 1</a> and sent data to quite a few researchers who <a href="/service/mailto:contact@adatascienti.st">emailed me</a>. I hope to see more people using this data to investigate how social networks can serve as echo chambers for conspiracies and misinformation.</p><p>Part 2 was supposed to be about misinformation websites shared within Parler, but that’s now in <a href="https://anonymousdata.medium.com/the-sources-of-misinformation-that-thrived-in-parler-cb740a98bd0d">part 3</a>. As I started analyzing external links from Parler, a trail of breadcrumbs led me to the bizarre campaign finances of Congress’ most controversial new members, <a href="https://en.wikipedia.org/wiki/Marjorie_Taylor_Greene">Marjorie Taylor Greene</a>.</p><h3>Parler’s link to WinRed</h3><p>A post on Parler, called a Parley, can have one or more links attached to it. Parler recorded the number of times each post was viewed, so I was able to visualize a breakdown of the most popular external domains.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_1kRTQfE56qeENhRySiyiQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HsiFFNeggaOVi716RwGgmg.png" /><figcaption>External domains weighted by number of impressions on posts that linked to them.</figcaption></figure><p>The most popular external domain <em>by impressions</em> is the fundraising organization <a href="https://winred.com/">secure.winred.com</a>, which distributes the money it raises to Republican political campaigns. There is clearly an advertising relationship between the Republican fundraising apparatus and Parler, confirmed by <a href="https://www.fec.gov/data/disbursements/?cycle=2020&amp;data_type=processed&amp;recipient_name=Parler&amp;two_year_transaction_period=2018&amp;two_year_transaction_period=2020">FEC filings</a>, as well as by the observation that measuring engagement in terms of comment count changes the distribution significantly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nw88yQCZ7PovaQZwDCNH8Q.png" /><figcaption>External domains weighted by number of comments on posts that linked to them.</figcaption></figure><p>According to <a href="https://www.fec.gov/data/disbursements/?cycle=2020&amp;data_type=processed&amp;recipient_name=Parler&amp;two_year_transaction_period=2018&amp;two_year_transaction_period=2020">FEC filings</a>, there was one super PAC and two political campaigns that used campaign funds to pay for advertising on Parler.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kWC_Ss3ynBh1c6SLrhu3_w.png" /></figure><ul><li>The campaign of <a href="https://en.wikipedia.org/wiki/Devin_Nunes">Devin Nunes</a> spent $1400 to advertise on Parler in mid-August. This is a <a href="https://www.opensecrets.org/members-of-congress/devin-nunes/summary?cid=N00007248">drop in the bucket</a> for the Nunes campaign, which raised over $26 million dollars.</li><li>The <a href="https://www.opensecrets.org/political-action-committees-pacs/stop-socialism-now-pac/C00755629/summary/2020">Stop Socialism Now</a> super PAC began advertising on Parler on November 9th, a few days after the US presidential election was called in favor of Joe Biden. On the 9th, they paid Parler $1000 in advertising fees. The following day they paid another $6000, and then another $8000 by the end of the week. The following week they paid $20,000 for advertising on Parler, spending a total of $41,075 in the month of November, all of it after the election.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cEiERayHtHOodaMWVJ9FCA.png" /><figcaption>The Stop Socialism Now super PAC is shown in blue. The contributions of this PAC were dwindled by that of the political campaign of Marjorie Taylor Greene.</figcaption></figure><ul><li>The largest and most bizarre source of advertising income for Parler during the 2020 election came from the campaign of <a href="https://en.wikipedia.org/wiki/Marjorie_Taylor_Greene">Marjorie Taylor Greene</a>, who spent $211,000 in less than one month. Greene had a guaranteed victory in her <a href="https://ballotpedia.org/Marjorie_Taylor_Greene">race for Georgia’s 14th Congressional District</a>, which leans heavily right. Greene’s campaign spent $4500 advertising on Parler over the week leading up to November 3rd, then proceeded to spend a whopping $206,500 over the remainder of the month, including a single payment of over $50,000. Her total spending on Parler represents around 10% of the <a href="https://www.opensecrets.org/races/candidates?cycle=2020&amp;id=GA14&amp;spec=N">total funds raised by her campaign</a>, and 83% of the total reported political expenditure on Parler for the 2020 election cycle.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*w3kpNx-2FoWYsB1J9eGfOA.png" /><figcaption>Parler took in $253,475 from political campaigns, nearly all of it after after the final vote was cast in the 2020 US elections. 83% came from the Greene campaign, 16% from a super PAC, and less than 1% from Devin Nunes.</figcaption></figure><h3>What was Greene advertising?</h3><p>Probably not her own campaign, considering she had already won in a landslide against an opponent who <a href="https://www.washingtonpost.com/nation/2020/10/17/kevin-van-ausdal-qanon-marjorie-greene-georgia/?arc404=true">effectively withdrew from the race</a>. Greene is rarely the subject of Parler posts that have boosted engagement (i.e. a higher than usual impression-to-comment ratio), and is hardly ever mentioned by Parler users. I cannot say anything with absolute certainty since the <a href="http://archiveteam.org/index.php?title=Parler">full archive of Parler</a> is not yet ready, but I find it highly unlikely my crawler would miss Greene’s political advertisements but catch all the others.</p><p>I can only speculate that this post from November 7th, which recorded over 8 million impressions, is related to the $9000 payment the Greene for Congress campaign <a href="https://docquery.fec.gov/cgi-bin/fecimg/?202012039339564123">made the day before</a>.</p><blockquote>Control of the Senate comes down to Georgia.<br>Do your part to defend the majority and keep Georgia red by splitting a donation between Kelly Loeffler and David Perdue today &gt;&gt; <a href="https://secure.winred.com/georgians-for-kelly/perdue-for-senate-defend-the-majority">Defend the Senate Majority</a></blockquote><p>On November 8th, the following post soliciting donations for the <a href="https://ballotpedia.org/Sean_Parnell_(Pennsylvania)">failed recount efforts of Sean Parnell in Pennsylvania</a> received over 9 million views, and likely performed quite well considering that Parler users were being continuously fed lies about election fraud.</p><blockquote>All eyes are on Pennsylvania, and our team is working day and night to keep this election fair.<br>If you can, please donate to our Election Defense Fund so we can make sure every vote is valid and legal.</blockquote><p>Sean Parnell lost the election for PA-17 by less than 10,000 votes, and was a vocal proponent of Donald Trump’s election fraud claims on <a href="https://twitter.com/SeanParnellUSA">Twitter</a>. The day after this post, the Greene campaign paid <a href="https://docquery.fec.gov/cgi-bin/fecimg/?202012039339564123">$50,500 to Parler</a>. The day after that, <a href="https://docquery.fec.gov/cgi-bin/fecimg/?202012039339564124">they paid $34,500</a>. By the end of the week, they would pay another $31,500. The Greene campaign effectively paid over $100,000 in a single week for the vague purpose of “digital advertising for fundraising” well after their candidate had already won her race.</p><h3>Isn’t this a campaign finance violation?</h3><p>The only way several of the sponsored advertisements on Parler in November could have been financed is by the campaign of Marjorie Taylor Greene. There is no other way for a Parley to organically reach millions of users unless promoted by a popular user like Sean Hannity or “echoed” by users, neither of which seems likely in this situation.</p><p>If my theory is correct, I believe it would be a <a href="https://www.fec.gov/help-candidates-and-committees/candidate-taking-receipts/support-other-campaigns/">violation of campaign finance laws</a> for the Greene campaign to put hundreds of thousands of dollars toward advertising for other Republican candidates without reporting an <a href="https://www.fec.gov/help-candidates-and-committees/filing-reports/in-kind-contributions/">in-kind contribution</a>. I believe that the campaign saw Parler as a fringe social network where neither the users nor operators would care to report an FEC violation against a Republican, especially given that Parler’s main investor is a <a href="http://cnn.com/2020/11/15/media/rebekah-mercer-parler/index.html">prominent conservative donor</a>.</p><p>Am I the only one who finds it strange that a candidate with a guaranteed win in her election decided to spend over $200,000 on political advertising <em>after</em> the final ballot was cast, barely made a scratch on the social network where she single-handedly contributed 83% of political advertising revenue, and didn’t have a <a href="https://docquery.fec.gov/cgi-bin/forms/C00708289/1408937/">single line item on her campaign receipts</a> indicating the money was dispensed to another campaign?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*p1rF_wbExbn6apzZDSmKOA.jpeg" /><figcaption>Greene has already become famous for her rambunctious statements after only a week in Congress.</figcaption></figure><p>It also strikes me as unethical that Greene’s construction company <a href="https://www.sba.com/ppp-funded-companies/georgia/taylor-commercial-inc-198827">received a PPP loan of up to $350,000</a>, even though she had the funds to <a href="https://docquery.fec.gov/cgi-bin/forms/C00708289/1408937/">loan $700,000 to her campaign</a> and to blow over $200,000 of that money on digital advertising that she didn’t need for herself.</p><h3>The super PAC also seems shady</h3><p>The <a href="https://www.opensecrets.org/political-action-committees-pacs/stop-socialism-now-pac/C00755629/summary/2020">Stop Socialism Now super PAC</a> is funded by a handful of donors contributing $500 — $5000, and one big-money donor who lives within Greene’s congressional district contributing $125,000 (see the <a href="https://docquery.fec.gov/pdf/849/202012019337372849/202012019337372849.pdf#navpanes=0">FEC filing</a>).</p><p>The SSN super PAC fills out the “Purpose of Expenditure” field for Facebook advertising expenses with the specific candidate being supported or opposed by their advertising campaigns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/753/1*OUYF6QB-qPHIka1EYUeWZg.png" /></figure><p>For every disbursement made to Parler, they write the exact same vague purpose that the campaign of Marjorie Taylor Greene uses — “Digital advertising for fundraising”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/737/1*H6OSIGlx4m02gMCU8eDAhg.png" /><figcaption>When this super PAC dispenses funds to Parler, they do it in the exact same vague way that the Greene campaign does.</figcaption></figure><p>This super PAC is managed by Jason Boles of <a href="http://rtastrategy.com/">RTA Strategy</a>, a political consulting firm that manages various PACs and political finance machines. When you go to RTA Strategy’s website, the first thing you will read on the splash page is shown below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gfrPL2bpAcPmj3Wu-MPqhA.png" /><figcaption>You can’t make this shit up.</figcaption></figure><h3>What’s next</h3><p>I was planning to make part 2 about the misinformation sites that are widely shared on Parler. However, within minutes of starting my analysis on external links, I noticed the strange distribution of links to winred.com, which led me to look up FEC filings, which led me to learn everything I possibly could about Marjorie Taylor Greene. Greene is a known <a href="https://www.npr.org/2020/08/12/901628541/qanon-supporter-who-made-bigoted-videos-wins-ga-primary-likely-heading-to-congre">QAnon supporter</a>, and months of scraping Parler has left me with a pretty good understanding of how their crazy theories work. You pull one thread here, another thread there, and pretty soon you are making a nonsensical argument for how Bill Gates started the coronavirus. I am legitimately worried I might be doing the same thing, seeing patterns where none exist. I will hopefully have more to write on this when two <a href="https://www.foia.gov/">FOIA requests</a> get back to me.</p><p>In <a href="https://anonymousdata.medium.com/the-sources-of-misinformation-that-thrived-in-parler-cb740a98bd0d">part 3</a>, I proceed with my original plan of writing about how most of the links to disinformation sites on Parler are hosted by big tech companies, and hopefully provide a compelling argument as to why banning Parler is a futile exercise in regulating the Internet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*s4AbMczyd5tH8W8DJPz2rA.png" /><figcaption>Most of the COVID disinformation that circulated on Parler is hosted on Youtube, owned by Google.</figcaption></figure><p><a href="https://twitter.com/anonymousdata_">Follow me on Twitter</a> to ask me questions or email me at <a href="http://contact@adatascienti.st/">contact@adatascienti.st</a>.I would be happy to send you my data for your own research, and can let you know when the complete archive of Parler from <a href="http://archiveteam.org/index.php?title=Parler">archiveteam</a> is ready.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ee6c526c1d44" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Violent hashtag frequencies in Parler]]></title>
            <link>https://medium.com/swlh/violent-hashtag-frequencies-in-parler-eddab2871b66?source=rss-9aad9a2ae2dd------2</link>
            <guid isPermaLink="false">https://medium.com/p/eddab2871b66</guid>
            <category><![CDATA[right-wing-extremism]]></category>
            <category><![CDATA[scraping]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[parler]]></category>
            <category><![CDATA[hacking]]></category>
            <dc:creator><![CDATA[a data scientist]]></dc:creator>
            <pubDate>Tue, 12 Jan 2021 09:02:52 GMT</pubDate>
            <atom:updated>2021-03-12T00:01:32.488Z</atom:updated>
            <content:encoded><![CDATA[<p>I ran a <a href="https://en.wikipedia.org/wiki/Web_crawler">crawler</a> to collect hundreds of millions of posts on <a href="https://en.wikipedia.org/wiki/Parler">Parler</a> before <a href="https://www.buzzfeednews.com/article/johnpaczkowski/amazon-parler-aws">Amazon shut them down</a> for hosting hate speech, conspiracy theories, threats of violence, and misinformation.</p><p>Users could only search posts on Parler (called “Parleys”) by hashtags, not by the text content within the post itself. This led users to liberally apply hashtags to any Parley that they wished to disseminate widely, providing valuable metadata to researchers like myself.</p><p>For example, here is a random Parley from January 6th in Washington D.C.</p><blockquote>#trump had to allow #mikepence to sign his own death warrant…..<br>#TraitorPence #GITMO #firingsquad</blockquote><blockquote>#ALLorNOTHING #2theEND #wildDC #magarally #wildprotest<br>#saveamericarally #trumprally #dc #ArrestPence #traitorVP</blockquote><blockquote>#JAN6 the new #911….. #makethempay #retribution</blockquote><p>This is the first part of a multi-part series where I analyze the data I collected from Parler. In the next part, I use this data along with FEC filings to make a case for why Georgia representative <a href="https://anonymousdata.medium.com/the-political-campaign-that-put-a-quarter-million-dollars-into-parler-after-the-election-ee6c526c1d44">Marjorie Taylor Greene may have committed campaign finance violations</a>.</p><p><strong>As of March 10th, </strong><a href="https://parler.adatascienti.st/"><strong>anyone can search the full dataset</strong></a><strong> and download my full dataset, with all archived pictures and videos from Parler’s CDN. Thank you to all the incredible people on the Internet who helped make this possible.</strong></p><h3>Hanging politicians goes viral on Parler</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Oqc152Ud7rDDchNIzc-tQg.png" /><figcaption>Relative frequency of hashtags that call for execution by hanging — e.g. “pelosi4gallows”, “hangtraitors”, “bringbackpublicexecutions”, etc. Yes, those are real hashtags that were widely used.</figcaption></figure><p>The <a href="https://parler.adatascienti.st/search?q=hang+them+for+treason">notion of murdering politicians</a> was discussed on Parler for months before the 2020 US election, often in response to perceived injustices perpetrated by that politician. The frequency of hashtags involving hanging/execution more than doubles after the 2020 US election (blue vertical line above), more than triples on the day of the Electoral College vote (green line), and is an order of magnitude higher on the day of the Capitol riot (red line).</p><p>Parler may have allowed the idea of attacking and hanging politicians to spread virally on the platform, most likely by negligence. It is my opinion that the notion of hanging, which was gaining wider acceptance in online discourse, manifested itself in the gallows erected outside the US Capitol on January 6th. I believe the gallows are a reference to the white supremacist literature <a href="https://en.wikipedia.org/wiki/The_Turner_Diaries">The Turner Diaries</a>, in which a “Day of the Rope” features politicians hung en masse.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*n3ibwyoH-d83I5Dc_1DhJA.jpeg" /><figcaption>A reference to the “Day of the Rope” where dissenters from white supremacist groupthink are systematically murdered, followed in the story by violent reprisals against non-whites and nuclear armageddon.</figcaption></figure><p>Below is a word cloud of hashtags associated with posts that explicitly mention murdering <em>specific</em> politicians (Nancy Pelosi, Joe Biden, AOC, and even lesser-known Democrats like Eric Swalwell).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*472O3BsdsVZ_58hHMBnk7w.png" /><figcaption>Hashtags that show up more frequently have a larger font, and it should be no surprise that many of the 45th president’s campaign slogans are so closely associated.</figcaption></figure><p>In the graph below, hashtags are connected and placed closer together when they are frequently mentioned together. Tags are grouped into communities that are frequently mentioned together, and edges span across communities to indicate association between various topics of discussion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Copm4ROVSk6AVGO_0Zi48w.png" /><figcaption>Visualizing discussion topics on Parler and how various discussions are linked by analyzing hashtag correlation.</figcaption></figure><h3>Exploring the dataset</h3><p>I use the <a href="https://wolfram.com/language">Wolfram Language</a> for analysis, which allows me to sample and search the dataset around particular topics of discussion. As I randomly perused based on hashtags in the discussion graph above, I realized how difficult moderating this kind of conversation would be. For example, consider the post below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FY8zjLcwY1yCB-xWwg_UDA.png" /></figure><p>The author of the post is sharing a book on guerilla warfare that is <a href="https://www.amazon.com/Fry-Brain-Sniping-Guerrilla-Warfare/dp/0971413398">available on Amazon</a>. I can look up the post author by their user ID, and from their other posts and comments it is obvious that they are trying to incite guerilla warriors to attack the “deep state”. Amazon is wise to not provide a platform for this information to disseminate digitally, but is it simply pushing people back to the printing press?</p><h3>The next viral hashtag</h3><p>In the aftermath of the January 6th insurrection, there was one particular class of hashtags that appeared to be mentioned more often.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lgu9uutWqQoGBNk-rMHHlQ.png" /><figcaption>The use of hashtags calling for civil war <em>increased</em> by around 30% after the failed insurrection at the Capitol.</figcaption></figure><p>The hashtags related to civil war are frequently associated with the <a href="https://en.wikipedia.org/wiki/Boogaloo_movement">Boogaloo movement</a> and the campaign of Donald Trump. It is possible we will see the real-world effects of this radicalization within the coming months.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gooN7pmtrEFriRU8RF7r9w.png" /><figcaption>Using the Wolfram Language to visualize relative frequency of hashtags with the given keywords. You can see toward the end that posts about hanging politicians are being replaced by posts calling for civil war.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4Jry4fG04JqEI2-kRqep0w.png" /><figcaption>Community associations of hashtags — these groups of tags are frequently mentioned together. Yellow and green are references to the Boogaloo movement, purple hashtags typically involve violence and anti-establishment topics, while red tags are typically associated with Donald Trump.</figcaption></figure><p>In a world of 24-hour news cycles and the proliferation of online disinformation, hashtags are ostensibly correlated to whatever is popular on the Internet at a particular time. On Parler, most users would just “echo” the posts of others — the Parler equivalent of a retweet.</p><p>The hash tags and post IDs allow <em>misinformation</em> to be “contact traced” across the network. In a future post, I will share my ideas on ways to measure the contagiousness of an idea and to quantify how radicalized a user has become.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lnOmZxY-iHpAbrPnA60OfA.png" /><figcaption>Word cloud for hashtags on posts that explicitly call for civil war in America.</figcaption></figure><h3>The QAnon conspiracy</h3><p>There are a substantial number of Americans (including <a href="https://www.rollcall.com/2020/11/05/qanon-goes-to-washington-two-supporters-win-seats-in-congress/">two sitting US congresswomen</a>) who believe there is a <a href="https://en.wikipedia.org/wiki/QAnon">vast global conspiracy by a cabal of Satan-worshipping pedophiles to consolidate power</a>, and that Donald Trump is secretly trying to bring them down. One distinct feature of the QAnon conspiracy is how new layers of rationalization are continuously added through widely disseminated videos — unlike the <a href="https://en.wikipedia.org/wiki/Moon_landing_conspiracy_theories">moon landing conspiracies</a>, QAnon’s purported conspiracy is ongoing, and continuously adapting to current events. As the theory evolves, it provides opportunities for anyone on the Internet to <a href="https://en.wikipedia.org/wiki/Pizzagate_conspiracy_theory">insert more dangerous ideas into it</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C8UhToSzAEYmQCTV3GISOA.png" /><figcaption>QAnon posts on Parler are frequently associated with COVID denial and a belief that Bill Gates/George Soros orchestrated the pandemic, with the “evidence” often presented in poorly-edited videos.</figcaption></figure><p>This is a popular QAnon post taken from a random sample of the dataset. It has been viewed over 181,000 times, reposted over a thousand times, upvoted over a thousand times, and has 319 comments attached to it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y0lP_7GW4WMrSGEfKXeDpg.png" /><figcaption>George Soros is a billionaire investor, Holocaust survivor, and well-known financier of liberal causes.</figcaption></figure><h3>What’s next</h3><p>You can now read part 2 in which the data indicates that a political campaign and a super PAC <a href="https://anonymousdata.medium.com/the-political-campaign-that-put-a-quarter-million-dollars-into-parler-after-the-election-ee6c526c1d44">shoveled over a quarter million dollars into Parler</a> in the span of one month <em>after the election</em>.</p><p>The <a href="https://www.reddit.com/r/ParlerWatch/comments/kuqvs3/all_parler_user_data_is_being_downloaded_as_we/">Internet Archive</a> is putting together a far more comprehensive dataset of posts, images, comments, users, and media from Parler which I recommend to data scientists looking to take up this project. My future plans for this project are:</p><ul><li>Perform NLP and sentiment analysis on the comments that users leave on posts, and identify which hashtags are associated with which sentiments</li><li>Analysis of disinformation sites, the rate at which they propagate across the Parler social network, and the hashtags most frequently associated with each one</li><li>Quantifying the radicalization of a user by showing a statistically-significant trend toward more violent ideas and hashtags</li><li>Clean up the data and make it publicly accessible to researchers</li></ul><p>You can contact me about this post and to request access to the dataset by <a href="https://twitter.com/anonymousdata_">messaging me on Twitter</a> or emailing <a href="/service/mailto:contact@adatascienti.st">contact@adatascienti.st</a>. I am trying to stay anonymous so these stories aren’t about me — they should be about the data. If you read anything else I write and find a partisan lean to it, please call me out on it, as it defeats the whole point of writing these if I’m just speaking to an echo chamber of my own.</p><h3>Please don’t make the problem worse</h3><p>I do <strong>not</strong> think the average Parler user deserves to be victimized, called a fascist, or have their privacy taken from them. I will not be taking part in anything like that.</p><p>The people who committed crimes should be treated as criminals. However, there were some pretty sociopathic people on Parler who need compassion and possibly mental help, not further provocation. I believe we need to study how technology enabled a group of people to radicalize each other and reinforced an echo chamber where conspiracies and racism could flourish.</p><blockquote>Disclaimer: since Parler has never enforced their Terms of Service with regards to violent speech, I slept fine at night ignoring their ToS with regards to automated scraping of the site.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eddab2871b66" width="1" height="1" alt=""><hr><p><a href="https://medium.com/swlh/violent-hashtag-frequencies-in-parler-eddab2871b66">Violent hashtag frequencies in Parler</a> was originally published in <a href="https://medium.com/swlh">The Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>