<?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 R. Michael Alvarez on Medium]]></title>
        <description><![CDATA[Stories by R. Michael Alvarez on Medium]]></description>
        <link>https://medium.com/@rmichaelalvarez?source=rss-7e735bbab7be------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*2FnmcQnW9Y6lNetBpBKAPw.jpeg</url>
            <title>Stories by R. Michael Alvarez on Medium</title>
            <link>https://medium.com/@rmichaelalvarez?source=rss-7e735bbab7be------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 11:27:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rmichaelalvarez/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[Five Important Social Media Trends in 2021]]></title>
            <link>https://medium.com/trustworthy-social-media/five-important-social-media-trends-in-2021-78862e71d51d?source=rss-7e735bbab7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/78862e71d51d</guid>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[2021-trends]]></category>
            <category><![CDATA[twitter]]></category>
            <dc:creator><![CDATA[R. Michael Alvarez]]></dc:creator>
            <pubDate>Wed, 30 Dec 2020 19:33:42 GMT</pubDate>
            <atom:updated>2020-12-30T19:33:42.546Z</atom:updated>
            <content:encoded><![CDATA[<p>What are the five important social media trends that the research community needs to pay attention to in 2021?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*lBYVUoqXY_FVjy1Y7cblYA.jpeg" /></figure><h3>1) New Regulatory and Legal Actions.</h3><p>We’ve all heard <a href="https://www.cnn.com/2020/12/10/tech/facebook-antitrust-lawsuit-analysis/index.html">the buzz about breaking up Facebook</a>, and calls for more regulation of social media and online sales giants like Google and Amazon. There’s no doubt that in the United States, with a new Democratic presidential administration taking office in late January, that we’ll hear more about regulating social media platforms and generating a more competitive social media environment.</p><p>While I’m not thinking that there will be big changes in the regulatory environment for social media platforms in 2021, with lawsuits and scrutiny of their business practices, I think that Google, Facebook, Twitter, and Amazon might engage in some (needed) corporate self-reflection, and that they might themselves take some steps to alleviate some of the political pressure on their business models.</p><p>One change in 2021 might be that social media platforms, Facebook in particular, will be more open and transparent about their platforms, and that as part of that movement, will provide researchers with improved access to data from their platform (<a href="https://medium.com/trustworthy-social-media/misinformation-on-facebook-aacada0d2f71">something that I wrote about recently in another Trustworthy Social Media post</a>). Twitter will be taking a number of steps in 2021 to give academics and researchers improved access to data from their platform, and that also might put some pressure on the other social media and online commerce giants to do the same.</p><p>I also suspect that we will see more efforts to push the social media giants to better police the use of their platforms. For example, doing a better and more transparent job of detecting misinformation, eliminating trolling and harassment, and generally making the use of their platforms a positive and informative experience, may improve in 2021. Providing users and researchers with more information about how they are enforcing their policies might improve in 2021. Researchers will want to stay well-informed about these efforts, and about <a href="https://medium.com/trustworthy-social-media/finding-social-media-trolls-with-dynamic-keyword-selection-b6754c8aefef">research that is being done to detect trolling, harassment, and misinformation online.</a></p><p>Finally, I wouldn’t be surprised to see some of the states take the lead in pushing for more regulation of social media platforms, in particular states (like California) that have the initiative and referendum process. <a href="https://www.jstor.org/stable/42864374">Political science researchers have shown that strategic politicians use the initiative and referendum process to put issues like these in front of voters,</a> and I would not be surprised if we hear discussion of potential ballot measures to further regulate social media platforms in 2021.</p><h3>2) Increasing Use of Video Content.</h3><p>It doesn’t take a Ph.D. researcher to tell you that social media platforms are seeing more and more video content (and that they are giving their users improved tools to post and edit video content). While the research community has made great advances in the analysis of the text information generated by social media platforms, we still lag behind the user community in our tools for collecting, storing, and analyzing the vast amount of video content that is being shared today on social media platforms.</p><p>By not analyzing video social media content, the research community is really missing out on a lot of important information. A social media post might have a short snippet of text, with an accompanying video, where the text does not adequately describe the details of the content in the video. Or that text snippet might be phrased in a way that leads natural-language processing methods astray, and we actually don’t get the correct interpretation of the content because we ignore the video.</p><p>The good news is that researchers are actively working to develop tools and methods that allow for the quick and easy analysis of images and videos; an excellent introduction is in this book, <a href="https://www.cambridge.org/core/elements/images-as-data-for-social-science-research/0376EE8A7A21F5B47FC4EC24DF07EFE9">Images as Data for Social Science Research</a> (recently published in a book series that I co-edit). While our methods for analysis of video social media content are lagging the dissemination of video content, I am sure that in 2021 we will see an increased focus on the development of tools for analysis of video content by researchers. I also anticipate that we see an increase in the quantity of interesting research papers published in 2021 that use data from social media video content.</p><h3>3) Now You See It, Now You Don’t.</h3><p>One issue that might be troubling for researchers is the trend towards the increased use of so-called “ephemeral content.” This is social media content that is posted but which automatically disappears after some period of time, and it’s becoming increasingly popular (for example, Snapchat and Instagram).</p><p>In 2021, we will see social media platforms provide their users with more tools to post transitory content. And I suspect that we will also see that users will take advantage of these tools. These trends pose problems for many types of research that are conducted using social media data, especially those that collect or scrape social media posts days, weeks, or months after they were originally published.</p><p>If more ephemeral content is posted in 2021, researchers might increasingly be missing out on social media platform conversations. If users take advantage of these tools for posting content that quickly disappears, and use them for certain types of conversations, this might introduce different types of biases in social media datasets and analyses. Researchers will need to follow this trend carefully, and in some cases, rethink their data collection methods. And researchers might need to start to think about how to detect and correct for these types of biases in their analyses of datasets where some content might have been self-censored.</p><h3>4) New Social Media Platforms?</h3><p>Don’t get me wrong, Facebook, Twitter, Instagram, and the other big social media platforms won’t disappear in 2021. But I think that we will continue to see new social media platforms emerge in 2021 (after all, I don’t think that anyone saw that TikTok would become so popular, and so controversial, in 2020!).</p><p>While new social media platforms won’t replace their bigger rivals in 2021, I think that we will see the continued emergence of new social media platforms in 2021. Researchers should pay careful attention to these new entrants, because like TikTok they will likely focus on trying to gain user shares quickly among very specific demographics, with very specific types of applications. If many members of specific demographic groups suddenly shift their social media from Facebook or Twitter to a new platform, that again produces potential biases in our data. In 2021 researchers pay more attention to these new social media platforms, like TikTok, which haven’t received a great deal of attention in the research literature.</p><h3>5) New Tools for Analysis of Social Media Data.</h3><p>Finally, I’m excited and confident that we will see many new tools and methods in 2021 for the analysis of social media. The research community has made great advances in the development of tools and methods for analyzing the text in social media data, and we will see the continuation of this in 2021. I think we will see text and natural language processing tools that will be faster and that will work with data at larger scales; our Trustworthy Social Media group is working on some new topic models methods in very large scale text datasets, for example.</p><p>Also, as I discussed earlier, there’s going to be a lot more development of tools for the collection and analysis of images and video. With so much social media content today including images and video, researchers will need to analyze those content alongside the accompanying text, which will complicate the computational tasks of working with social media data.</p><p>Finally, I hope that in 2021 we also see that researchers develop and release more tools for the collection of social media data. <a href="https://medium.com/trustworthy-social-media/how-do-we-monitor-social-media-bdd6e9dc6c3c">Our group has released the code for our cloud-based social media collection infrastructure</a>, and I expect that we will see other research groups share similar code and tools in 2021.</p><h3>Conclusion</h3><p>In conclusion, 2021 is going to be a challenging and exciting time to be collecting and analyzing social media data. With so much of the world’s communications shifting to social media platforms, with the emergence of new social media tools, and with the shift towards video content, we will have a lot of work in 2021 to continue our research towards building more Trustworthy Social Media. I’m confident that 2021 is going to see a great deal of innovating and exciting research and scientific developments!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=78862e71d51d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/trustworthy-social-media/five-important-social-media-trends-in-2021-78862e71d51d">Five Important Social Media Trends in 2021</a> was originally published in <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Misinformation on Facebook]]></title>
            <link>https://medium.com/trustworthy-social-media/misinformation-on-facebook-aacada0d2f71?source=rss-7e735bbab7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/aacada0d2f71</guid>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[misinformation]]></category>
            <category><![CDATA[twitter]]></category>
            <dc:creator><![CDATA[R. Michael Alvarez]]></dc:creator>
            <pubDate>Tue, 29 Dec 2020 18:08:53 GMT</pubDate>
            <atom:updated>2020-12-29T18:08:53.953Z</atom:updated>
            <content:encoded><![CDATA[<p>As researchers, we do not know enough about misinformation on Facebook.</p><p>Why? Because Facebook does not give researchers easy access to content from their platform.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*qCKfJWn-gcO8Ba8gLbbpxw.jpeg" /></figure><p>Everyone hears stories about the rampant misinformation and negativity on Facebook. Recently, Charlie Warzel wrote an interesting article about this trend in the New York Times titled <a href="https://www.nytimes.com/2020/11/24/opinion/facebook-disinformation-boomers.html?searchResultPosition=3">“What Facebook Fed the Baby Boomers.”</a> In the article, he discusses how he worked with two baby boomers’ to access their Facebook feeds, and describes just how much toxic and misleading content the subjects were exposed to. Warzel summarizes the experience of one of the two baby boomers as “an information hellscape.”</p><p>It’s an interesting read, but a question remains: why did Warzel take the approach of following just two users? He writes in the article:</p><blockquote>“Despite Facebook’s reputation as a leading source for conspiracy theories and misinformation, what goes on in most average Americans’ news feeds is nearly impossible for outsiders to observe. Tools like CrowdTangle, which track “engagements” with social media posts, are the best available means to understand what is popular on the platform, though Facebook (which owns the CrowdTangle) argues that CrowdTangle is not a reliable indicator for how many people saw a post.”</blockquote><p>Unfortunately, the approach used by Warzel is one of the few ways that a researcher can get access to the sort of information that flows through most Facebook accounts. But while Warzel’s approach is interesting (and makes for a great read), it does not provide timely and systematic data that researchers need to understand how much misinformation, harassment, trolling, and hate speech exists on Facebook — nor does this approach give us the data to build tools that can detect and perhaps prevent misinformation on Facebook.</p><p>Of course, you might be thinking, can’t a researcher build a scraper that could quickly collect a great deal of Facebook information? The answer is yes, but that could violate the current <a href="https://www.facebook.com/legal/terms">Facebook Terms of Service</a>. The TOS states: “You many not access or collect data from our Products using automated means (without our prior permission) or attempt to access data you do not have permission to access.” So an effort by an independent researcher to scrape data at scale seems out of the question.</p><p>Of course, many of us know the story that led to these restrictions on data sharing with researchers. In the wake of the Cambridge Analytica scandal (the details of which have been recounted numerous times, for a good overview I’d suggest reading <a href="https://www.amazon.com/Mindf-Cambridge-Analytica-Break-America/dp/1984854631/ref=sr_1_1?dchild=1&amp;keywords=christopher+wylie&amp;qid=1609265191&amp;s=books&amp;sr=1-1">Christopher Wylie’s book Mindf*ck</a>), Facebook essentially shut down research access to data from their platform. In the two and a half years since then, the research community has had very limited access to large, timely, and researcher-oriented datasets from Facebook. Facebook essentially cut off all direct research access to content from their platform.</p><p>Today, Facebook makes available a limited selection of curated datasets for academic research on their <a href="https://research.fb.com/data/">Facebook Data for Independent Research</a> site, but that’s a far cry from allowing researchers to collect their own datasets. There’s also the <a href="https://socialscience.one/">Social Science One</a> initiative, but it involves an opaque process to request access to curated datasets (right now it says that one can apply for access by sending a request on the proposals section of a website that doesn’t seem to exist at this time. I’m happy to be corrected about that if I’m wrong).</p><p>But this is not enough. Facebook has billions of users worldwide, and stories like Warzel’s imply that Facebook may have become one of the primary mechanisms for misinformation and vitriol to spread online. Facebook needs to start restoring direct research access to their platform, following the lead of their competitor, Twitter. Not only has Twitter given researchers direct access to content from their platform, <a href="https://developer.twitter.com/en/products/twitter-api/early-access/guide">but they have also recently announced a series of steps they will take in 2021 to provide qualified researchers even greater access to content from their platform.</a> And by this I mean opening an API that allows research queries for current Facebook content, at the sort of scale that would allow researchers to study the flow of information (and misinformation) on Facebook.</p><p>What are we missing by not having direct and unimpeded access to current content on Facebook? As one example, if our group wanted to analyze the spread of misinformation in the 2020 presidential election on Facebook, in the same way that tracked on monitored these issues using Twitter data, we simply could not replicate this methodology using the Facebook platform. <a href="https://medium.com/trustworthy-social-media/what-are-people-tweeting-about-the-2020-presidential-election-f3dc96ff838b">The word cloud below comes from our Twitter election monitor</a>, and it shows the type of data that we collect from Twitter (which isn’t readily available from Facebook). Nor can we dig deeply into datasets on the 2020 presidential election to detect what may be small but influential networks that spread misinformation on Facebook. Thus we really cannot understand the full ecosystem of misinformation online, because we can’t easily and independently get a complete dataset on what is happening on Facebook.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*1Cq7AkvyXNdGo8sQMja6MQ.jpeg" /><figcaption>From the Twitter election monitor at monitoringtheelection.us</figcaption></figure><p>There’s even the chance that since those who deliberately spread falsehood online know that the research community has little ability to track their misinformation campaigns on Facebook, they might be more inclined to use Facebook for their activities.</p><p>In order for research groups like ours to help companies like Facebook build a truly <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a>, we need timely and detailed access to data from their platform. We hope that in 2021, Facebook provides better opportunities for researchers to collect the data they need for their studies, to gather data at the scale necessary for in-depth analysis of content and spread of misinformation, and to be able to gather the data without Facebook being able to restrict the types of research that use content from their platform.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aacada0d2f71" width="1" height="1" alt=""><hr><p><a href="https://medium.com/trustworthy-social-media/misinformation-on-facebook-aacada0d2f71">Misinformation on Facebook</a> was originally published in <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Was YouTube Too Slow To Remove Election Fraud Content?]]></title>
            <link>https://medium.com/trustworthy-social-media/thanks-youtube-but-you-might-be-too-late-7fe559aff069?source=rss-7e735bbab7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/7fe559aff069</guid>
            <category><![CDATA[election-fraud]]></category>
            <category><![CDATA[youtube]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[misinformation]]></category>
            <category><![CDATA[election-2020]]></category>
            <dc:creator><![CDATA[R. Michael Alvarez]]></dc:creator>
            <pubDate>Thu, 10 Dec 2020 22:02:58 GMT</pubDate>
            <atom:updated>2020-12-13T16:24:44.492Z</atom:updated>
            <content:encoded><![CDATA[<p>Yesterday, <a href="https://blog.youtube/news-and-events/supporting-the-2020-us-election/">on their official blog</a>, YouTube announced that they will (now) begin to take down “any piece of content uploaded today (or anytime after) that misleads people by alleging that widespread fraud or errors changed the outcome of the 2020 U.S. presidential election …”. This is a good step, but there’s a strong chance that by not acting earlier to take this content down, YouTube (and other social media platforms) might have fueled voter distrust and lack of confidence, especially among Trump voters and Republicans.</p><p>YouTube justified this change in policy by noting that the safe harbor deadline for the election had passed, and that a sufficiently large number of states have now certified their presidential election results. Whether in YouTube’s opinion that means that the claims of election fraud or errors are now irrelevant, or if they believe the fraud claims do not have factual support, isn’t clear from their post. And while the post notes many of the proactive steps that YouTube took during the election (they mention deleting over 8,000 channels and “thousands of harmful and misleading election-related videos”) are important, that they still allowed misleading content about election fraud with little factual basis to remain on their platform might have lasting consequences for how confident American voters are in the election process. Finally, there are important distinctions between YouTube’s new policy, and the policies that were implemented during the 2020 Presidential election by Facebook and Twitter, but that’s really a topic best taken up in a later story that can really dig into those differences.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*D16I_Dnx7nhPN-S_WIfgAA.jpeg" /></figure><p><a href="https://www.nadamscohen.com/">Nicholas Adams-Cohen</a>, a member of our research group, wrote a fascinating piece on Medium last week, <a href="https://medium.com/trustworthy-social-media/suppression-or-manipulation-8ad54f4def22">“Suppression or Manipulation? Comparing Perspectives on US Election Security Between Biden and Trump Voters.”</a> He shows a really important figure in his piece, demonstrating the highly partisan split in the type of concerning information received on social media about election integrity during the election. In particular, Nicholas showed that Trump voters were much more likely than Biden voters to receive concerning social media information on voter fraud, voting machine tampering, duplicate voting, and non-citizen voting, many of the same themes that are no doubt in the content that YouTube will now not allow on its platform.</p><p>In a related report, the <a href="https://monitoringtheelection.us/">Monitoring the Election project</a> at Caltech released a <a href="https://static1.squarespace.com/static/5ace8a6b45776eba2e40cbee/t/5fb6d361f0ceb67888b03c1c/1605817186684/Voter_Confidence_2020_Survey_Report.pdf">research memo last week presenting new data on voter confidence in the immediate aftermath of the November 2020 election</a>. This report shows that there are very substantial gaps in voter confidence, in particular regarding whether the election was administered well nationally, between Republican and Democratic registered voters, and also between Trump and Biden supporters.</p><p>Remarkably, the <a href="https://monitoringtheelection.us/">Monitoring the Election</a> report showed that 44% of Trump voters said that they were not at all confident in election administration at the national level, and that another 26% of Trump voters were not too confident. But 48% of Biden voters were very confident in the nation’s election administration, and 39% were somewhat confident.</p><p>Correlation is not causation, and we need to be careful attributing cause-and-effect. But these data imply that since Trump supporters were more exposed to concerning social media claims about voter fraud, and since Trump voters lack confidence in the national administration of the election, that by not doing a better job removing misleading or false claims about voter fraud social media platforms like YouTube might have allowed the distribution of information that has fueled this large partisan gap in voter confidence.</p><p>Clearly we need to do a great deal of additional research to better connect social media content, exposure to that content, and attitudes and perceptions about the integrity of the 2020 presidential election. We need to give social media platforms better guidance, based on the science, about how their content may lead to problematic outcomes that could have lasting consequences for how much trust American voters have in their election process and voting infrastructure. Research groups like ours also need to do more analysis of the effects of these policies, stay tuned!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7fe559aff069" width="1" height="1" alt=""><hr><p><a href="https://medium.com/trustworthy-social-media/thanks-youtube-but-you-might-be-too-late-7fe559aff069">Was YouTube Too Slow To Remove Election Fraud Content?</a> was originally published in <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Do We Monitor Social Media?]]></title>
            <link>https://medium.com/trustworthy-social-media/how-do-we-monitor-social-media-bdd6e9dc6c3c?source=rss-7e735bbab7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/bdd6e9dc6c3c</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[twitter]]></category>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[R. Michael Alvarez]]></dc:creator>
            <pubDate>Tue, 24 Nov 2020 21:55:10 GMT</pubDate>
            <atom:updated>2020-11-30T20:54:18.422Z</atom:updated>
            <content:encoded><![CDATA[<p>One important step towards building tools to building a more trustworthy social media is to develop reliable methods for collecting social media data.</p><p>How do we monitor social media data, especially over long periods of time?</p><p>Our research group has been monitoring various topics on Twitter since 2014. For example, our <a href="https://rmichaelalvarez.github.io/twitter-monitors.html">election monitor </a>has run in the 2014, 2016, 2018, and 2020 US election cycles. Below is an example of a word cloud of tweets we have collected recently on election fraud.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ygJHIto4zA1GrTSN-ET2OA.jpeg" /><figcaption>Word Cloud of November 30, 2020 Tweets About Election Fraud</figcaption></figure><p>We’ve also used our monitor to collect conversations on Twitter over long periods of time on other topics, like immigration, gay marriage, abortion, and gun control. These data have been used by some members of our research group, for example, <a href="https://617c12b8-8dc2-4abb-b2ae-49d765ddb31e.filesusr.com/ugd/c31be4_56fa1dcfbed543dcb96438118b04083d.pdf">Nicholas Adams-Cohen’s work using one of these datasets to study how conversations about gay marriage on Twitter were altered by the landmark Obergefell v. Hodges decision.</a></p><p>Since 2014, we’ve been building various types of architectures to reliably and robustly collect data from Twitter’s public “streaming API” over long periods of time. We’ve found that many things can happen that can interfere with collecting Twitter data in real-time over periods of weeks, months, and even years. For example, sometimes when we have used local servers to collect these data, the servers fail; sometimes the local network fails, and in other situations, things happen that force the monitors to stop collecting data, sometimes at important periods of time.</p><p>Recently, to try to alleviate these problems, we’ve been moving our architectures to the cloud. While that might seem like an easy thing to do, in particular for readers who know how to write python code and who can work in the cloud, in all honesty, it’s harder than you might think!</p><p>Even better — we developed the architecture to collect, pre-process, and store our Twitter monitor data using three different cloud platforms: the Google Cloud Platform, Oracle Cloud, and Amazon Web Services. While the details vary by platform, we have built a generalized approach that others can use for their research using Twitter’s public streaming APIs (as well as other public streaming APIs).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WN5BVgqz64-OYEI_jkrJ0A.png" /></figure><p>Basically, our architecture uses this simple workflow with four different parts. In the first stage, a producer requests the social media data from a public API, say Tweets for certain election keywords (like our current election monitor). The producer then passes information in real-time to a data streaming for short-term storage, essentially a place where we can time-stamp the data and make sure that we are capturing it. The data consumer then grabs the data from the data stream, and at that point, can either pass them to other workflows for pre-processing (say labeling or geotagging) or it can send them directly to some storage solution. Sometimes the storage component is short-term cloud-storage, in other cases we secure the data in the cloud but also pass it to the research group using an accessible storage and sharing application (like Google Drive, Box, or Dropbox).</p><p>We’ve written a working paper that presents this architecture in more detail, and provides information about how we implement it using three different cloud computing systems. <a href="https://arxiv.org/abs/2005.02442v3">The paper, “Reliable and Efficient Long-Term Social Media Monitoring”, written by Jian Cao, Nicholas Adams-Cohen, and myself, is available as a preprint online.</a></p><p>We’re working to make the code used for these architectures available for others to use, and recently <a href="https://github.com/jian-frank-cao/MonitoringTwitter">Jian Cao (a postdoctoral scholar in our research group) put it on his GitHub repo</a>. We hope that our paper, and the availability of our code, encourage others to use these tools, to improve upon then, and to also share their methods and tools for collecting and analyzing long-term social media datasets.</p><p>We’re also working on a process for sharing much of the archival data that we have collected using these processes going back to 2014. More on that soon!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bdd6e9dc6c3c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/trustworthy-social-media/how-do-we-monitor-social-media-bdd6e9dc6c3c">How Do We Monitor Social Media?</a> was originally published in <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Is Trustworthy Social Media?]]></title>
            <link>https://medium.com/trustworthy-social-media/what-is-trustworthy-social-media-294886a3d1cf?source=rss-7e735bbab7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/294886a3d1cf</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[computer-science]]></category>
            <category><![CDATA[researchers]]></category>
            <category><![CDATA[introduction]]></category>
            <category><![CDATA[social-science]]></category>
            <dc:creator><![CDATA[R. Michael Alvarez]]></dc:creator>
            <pubDate>Sat, 14 Nov 2020 18:46:56 GMT</pubDate>
            <atom:updated>2020-11-14T21:34:57.118Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u0NfmqA0Ec_0JXiJ6npSMw.jpeg" /></figure><h3><strong>A statement by project leads Anima Anandkumar and Michael Alvarez</strong></h3><p><strong>Trustworthy Social Media</strong> is new publication from our group computer science and social science researchers, who are working together to return civility to discourse on social media, to help find and prevent online trolling and harassment, and to use AI to uncover misinformation and disinformation campaigns online. We also are working to show how we can use social media data productively for academic research.</p><p>We believe that social media trolling, harassment, and misinformation are some of the most important social, political, and economic issues of our time. Social media platforms like Facebook, Instagram, and Twitter have become important tools that people around the world us to communicate and get their news, they are widely used for e-commerce and advertising, and have become important mechanisms for politicians to use campaign for office and stay connected to their constituents.</p><p>Our research effort is a collaboration between the research groups of computer scientist <a href="http://tensorlab.cms.caltech.edu/users/anima/">Anima Anandkumar </a>and social scientist <a href="http://www.rmichaelalvarez.com/">Michael Alvarez</a>, both faculty members at the <a href="https://www.caltech.edu/">California Institute of Technology</a>. We have post-doctoral scholars, graduate students and undergraduates who have been working with us for over the past year to initiate and sustain a variety of related research efforts that all are aimed at producing a more trustworthy social media.</p><p>We plan on using this Medium publication to give readers some insight into our work-in-progress, to discuss the research that our team and others are conducting in this area, and to also discuss the policy and social implications of our work. We’ve built a <a href="https://anqiliu-ai.github.io/project/social/">project website</a> where we have links to the various components of our research project, our papers and other research products, and where we introduce the members of our research group.</p><p>Stay tuned to Trustworthy Social Media.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=294886a3d1cf" width="1" height="1" alt=""><hr><p><a href="https://medium.com/trustworthy-social-media/what-is-trustworthy-social-media-294886a3d1cf">What Is Trustworthy Social Media?</a> was originally published in <a href="https://medium.com/trustworthy-social-media">Trustworthy Social Media</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>