MarcoCivil: Visualizing the Civil Rights Framework for the Internet in Brazil Lorena Lucas Regattieri Fabio Malini Fabio Goveia LABIC-UFES LABIC-UFES LABIC-UFES regattie@ualberta.ca fabiomalini@gmail.com fabiogv@gmail.com Gabriel Herkenhoff Labic-UFES gabriel.herkenhoff@gmail.com ABSTRACT Civil Rights Framework for the Internet, journalism, data mining, In this paper, we map the controversy surrounding the Marco social network analysis, complex networks. Civil da Internet (Civil Framework for the Internet) in Brazil. Drawing on a Twitter dataset spanning from August 2012 to 1. INTRODUCTION December 2013, this study uses a series of methods of data After an intense debate in 2007-08, the Office of Legislative mining, processing, and information visualization to produce a Affairs of the Ministry of Justice, in partnership with the School historiography of collective actions related to the Marco Civil. of Law in Rio de Janeiro at the Getulio Vargas Foundation The MarcoCivil platform at the “Digital Culture” website creaed (FGV), initiated the collaborative construction of the proposal for initiatives to spread the discussions online: a Twitter profile a civil law framework for the Internet in October of 2009. The @MarcoCivil (run by the administrators of the platform) and the goal was to create legislation that defined "the legal responsibility MarcoCivil hashtag. To conduct the Marco Civil cartography we for providers and users for the content posted on the chose to work with the messages indexed to the MarcoCivil Internet…[and identified] measures to preserve and regulate the hashtag circulating on Twitter. In 2012 and 2013 Twitter became fundamental rights of Internet users, such as freedom of the online space in which cyber activists were most vocal. From expression and privacy".1 October 2012 to January 2013 , we collected about 21.997 tweets related to Marco Civil, it was then that we noticed the presence of The Civil Rights Framework for the Internet in Brazil opposes the a controversy and a diversity of points of view in dispute. News tendency to establish restrictions, convictions or bans on the use reports in Brazilian newspapers during the discussion, little took of the internet. The framework intended to determine clearly the into consideration the issues engendered in the struggle for rights and responsibilities regarding the use of digital media. The approval of the law. By demonstrating with graphs the dispute focus, therefore, is the establishment of a legislation ensuring between the different actors involved in this battle, we seek to rights, not a rule restricting freedoms. Between November 2009 contribute to the history of the approval of the Marco Civil. From and June 2010, the Marco Civil was developed through a uniquely telecommunications companies to politicians, our report show open public process that allowed all Brazilian Internet users an how history was made in the field of the internet human rights. opportunity to comment on its text. In the spirit of the bill’s substance, civil society was empowered to collaborate with policymakers in order to make the bill reflective of public interest Categories and Subject Descriptors and prioties. D.3.3 [Programming Languages]: Language Constructs and Features – abstract data types, polymorphism, control structures. An initial draft drawn by legislators was posted Cultural Digital, an open platform where the public could submit and review K.4.1 [Computer and Society]: Public Policy Issues – Ethics, suggested changes to the bill. Throughout an open debate, Marco Intellectual property rights, privacy, regulation. Civil received over two thousand comments from academics, civil G.2.2 [Numerical Anlaysis]: Graph Theory - Graph algorithms. society organizations, technical experts, and private individuals. Network problems. In 2011, the Marco Civil was submitted to Congress as Executive Bill 2126 and was given priority on the legislative agenda. Since then, the bill has become the subject of numerous controversies in General Terms the House of Representatives due to inflammatory issues such as Algorithms, Management, Measurement, Documentation, network neutrality, privacy, freedom of expression, and copyright. Performance, Design, Reliability, Experimentation, Human The Bill has made it onto the agenda of the House of Factors, Languages, Theory, Legal Aspects, Verification. 1 An English version of the bill is available at FGV Keywords http://direitorio.fgv.br/sites/direitorio.fgv.br/files/Marco%20Civ il%20-%20English%20Version%20sept2011.pdf Representatives eight times, but each time the vote has been ● Map the network of controversies on the #MarcoCivil; postponed due to the lack of agreement among Members about ● Perform a semantic analysis of the expressions, hashtags, and crucial points in the Marco Civil. controversial issues that circulated on Twitter under the Challenges in reaching an agreement have created an obstacle to #MarcoCivil hashtag. the consolidation of a national-level regulatory framework for the We centered our analyses around two distinct periods: Internet. Among other things, this immobility reveals a tension between the interests of businesses and the demands of civil ● July - December 2012: The Marco Civil bill enters the agenda society. Over the course of the bill’s legislative history, the of the House of Representatives telecommunications lobby and content industries have been the driving force behind significant changes to the text. During this ● July - December 2013: Discussions about the bill resume at the period, we have also witnessed a somewhat "schizophrenic" House of Representatives. dynamic take hold of policymaking efforts concerning the In our network visualization, we chose to plot the network of Internet. While the Ministry of Justice created an innovative retweets (RTs) that included the #MarcoCivil hashtag. Since RTs collaborative platform so that civil society could participate in the must be replicated by many individuals, RTs on Twitter indicate production of "The Bill of Rights for the Internet,” it also saw that a subject (represented by a hashtag) carried significant social broad mobilization around a bill that sought to combat all forms relevance. We extracted data directly from the Twitter API, which of crime on the Internet, especially financial crimes. Meanwhile, allowed us to capture and store about 20,000 tweets produced by the Parliament endeavored to focus on criminal laws as a almost 10,000 profiles monitored in 2012. foundational aspect of Internet regulation in the country. For each tweet, we were able to log the tweet text, date, origin and This strange situation persists today, as the copyright and destination of the tweet. The subsequent step after mining and telecommunications industries oppose free “peer to peer” processing is the data is the visualization of data. Using the open exchange and net neutrality. This can be explained, in part, by the source tool Gephi2, we sliced the data using different metrics, interests of public security forces, which after public protests in creating new graph visualizations for each metric. To support our June 2013 (strongly articulated by the civil society through social semantic analysis of the data, we analyzed 5137 tweets to identify networks) advocated establishing a longer required period for the the political position of each actor in the debate on #MarcoCivil; retention of private communications data that could support the the way Twitter profiles were expressing themselves in the investigation of crimes and "deviations". The situation was network; the intention of the message; the themes it touched upon; compounded in the wake of the Edward Snowden leaks revealing and the controversy. the National Security Agency (NSA) spying other countries through PRISM. This struck a chord for Brazilian President The second procedure was to analyze all the tweets, 21,000 in Dilma Roussef, who subsequent to the leaks, proposed an 2012 and 110,000 in 2013. For this, we used a data-mining tool amendment to the Marco Civil that would force foreign called NAR_T3, a python script developed within the Laboratory companies to host data on national servers. The proposal has of Studies on Image and Cyberculture (LABIC). The script proved highly controversial, due both, to the geopolitical provides the following outputs: implications it would carry and the technical complications it • Most repeated words and hashtags. could introduce. • Most replicated tweets. Within the approaval of the Marco Civil, the world turns the eyes to Brazil when it comes to Internet civil rights. The world • Word clouds and hashtags. celebrated the bill at the NETmundial – Global Multistakeholder • Co-occurring hashtags network. Meeting on the Future of Internet Governance and at Arena Participative. At the Arena, we had the presence of important • Most mentioned users. people discussing internet and human rights, such as Roy Singham (ThoughWorks), Julian Assange (Wikileaks) from the • Number of tweets per user. Ecuador Embassy, and Frank La Rue (ONU). The event that • Number of active users per day brought together representatives of governments and civil society in search of a letter of international principles for the Internet was After generating groups with Gephi, we extracted the profile considered the beginning of the process to discuss the internet names that built up each cluster in the network. When we policies in a global context. History was made, but it is crucial to processed the script with the "cluster_usernames" of each of the understand the path to the approval of the Marco Civil in order to groups, we obtained the same outputs, but now we could analyze comprehend the struggles involved in the fight for Internet human them by targeted group. This allowed us to investigate the unique rights. positions surrounding the controversy of each of the groups identified. 2. METHOD AND GOALS Latour[1] and Venturini’s[2] mapping controversies technique is successful method to trace digital data. It is broadly used in the communications field to map the debates around a specific object/event. This is the theoretical foundation guiding our 2 Gephi is an open-source software for visualizing and analyzing research; we used the cartography method to support us in the large networks graphs. Available at: http://gephi.org digging experience in the Twitter data. As an empirical template, 3 This script was created to parse tweets. It is available at Twitter served us for the purpose of: https://github.com/ufeslabic/parse-tweets 3. DISCUSSIONS volume of tweets eventually formed an interactive network with different common points of view on distinct aspects of the law 3.1 General Observations of Marco Civil (Figure 2). Network Dynamics In August 2012, when the Marco Civil entered the voting agenda at the House of Representatives, the politics of this power struggle overflowed into the virtual universe, particularly on social networks. This chart represents the high level of participation on Twitter, especially, the days in which the bill was expected to be voted at the House of Representatives. Figure 3. High rate Hashtag use with the hashtag #MarcoCivil. Figure 1. Number of Tweets per day with the hashtag #MarcoCivil on Twitter, from 21 August to 3 December 2012. With the vote imminent, activists, parliamentarians, lawyers, specialists, businessmen, intellectuals, artists, government ministers and even President Dilma Rousseff used social networks to produce a broad debate on the subject. The buzz over the Marco Civil quickly became one of the longest standing controversies in the recent history of Brazilian politics. The increasing rate of publication of tweets directly correlates with increased political debate around the subject. The closer the House of Representatives was to voting on the legislation, the more activity we saw on Twitter under the #MarcoCivil hashtag. The representatives found themselves facing pressure from a Figure 4. Word frequency within tweets mentioning the broad range of channels: social networks, emails, blogs, and hashtag #MarcoCivil online media. Some party websites even underwent DDoS attacks. Frequent use of the terms "vote" (votação) (6652), "postpones" Digital expression around the issue became a strategy for activists. (adia) (2065), "House"(câmara) (4941) and "bill of law" (projeto In many ways, these tactics exposed many politicians to public de lei) (2616)" suggested high levels of expectation that the bill judgment, affecting their image among voters. This strategy has would pass and a commitment, at least among a minority of users, proven to be a key measure to the movements connected to the to monitoring the long and tiresome journey of Marco Civil in the field of free culture and the most progressive deputies. Congress. The anxiety around the bill was highlighted by the intense correlation of the hashtag #marcocivil with the #MarcoCivilJá (#MarcoCivilNow). The word “neutrality” and the #neutrality hashtag can be seen often in the dataset (Figure 3 and 4) suggesting it was the most commonly discussed subject in interactions between members of Congress and users tweeting about #MarcoCivil. 3.2 Marco Civil in 2013: the network is polarized and the privacy debate gains attention In 2012, the difficult process in voting the bill 2126/2011, plus the numerous delays and changes in the course of the project, Figure 2. Number of unique users per day participating in the turned the social networks - notably Twitter - into a major publication of tweets with the hashtag #MarcoCivil. platform for discussion about the Marco Civil. Activists, experts and concerned individuals began to debate the issue, seeking to defend their perspectives and understand the significance of the From August to December 2012, heightened publicity around the bill for the future of the Internet in the country. But with the bill generated the mobilization of 16,072 different profiles, 22651 failure to reach an agreement and the start of the municipal tweets and 5640 retweets (Figure 2). A variety of profiles and the elections of 2012, the vote on the Marco Civil fell into oblivion, eventually being suspended. In June of 2013, two critical events links between the actors in the network and illustrates a force of affected the trajectory of the bill: Public uprisings throughout the attraction between them (a dynamic referred to as “gravity”.) As country and the first of the Snowden leaks. Protests over transit an individual, typically (thought not always) shares ideas with fare hikes, economic inequality and other “bread and butter” those, which he agrees, individuals with similar opinions share issues peaked in June, with some protesters referencing the bill content with and from each other, creating groups, which we call and making it part of their messaging, both on and offline. At the perspectives. There are four perspectives within the Marco Civil same time, some activists began to argue against the creation of network: the civil framework for the Internet, claiming that the Marco Civil • The purple network: individuals in favor of voting on the law was a ploy made by the government to restrict Internet freedom. (46.55% of the total network) This questioning came up in light of the numerous arrests of Facebook page administrators from groups opposed to the •The red network: individuals contrary to voting the bill government, in particular, Anonymous and Black Blocs [3]. Back (17.39%) then, videos from Anonymous began circulating claiming that Marco Civil was going to have the opposite effects: for them the • The green network: media outlets and profiles specialized in intentions of Marco Civil were to control online content. Thus, a law and civil rights (20.56%) trend of polarization emerged while some continued to promote • The yellow network: foreign organizations that generally the bill, despite changes in the text that weakened user protections supported the proposal for a regulatory framework (4.1% of total) in the face of copyright restrictions, others began voicing opposition to the bill, arguing that it would lead to greater Internet censorship. The perspective of media outlets came out exactly 4. CONCLUSIONS between these two groups, as news feeds reflected the arguments Our study suggests that the free digital culture activists are the of both sides. The emergence of groups that made radical critiques ones responsible for articulating the Marco Civil debates. Thus, of the Marco Civil represented a fundamental shift in the debate social networks i.e. Twitter, prove to be a rich environment for the on the subject. This change can be better understood when we open debate. This network has become a major strategy to undertake a semantic analysis of the network formed by these pressure the Brazilian Congress. In our study, we employed groups during this period of time. computer-assisted analysis through mining methods and data visualization in order to investigate our hypothesis. The outputs have proven that our hypothesis is correct, as our research displays several indications pointing to the centrality of the actions and pro-Marco Civil campaign coordinated by activists from Brazil and around the world. The days before voting on the Marco Civil by the House of Representatives were periods when Twitter profiles became highly mobilized in order to debate and press the Parliament on the approval (or not) of the Marco Civil. This demonstrates that the community formed around the hashtags remained attentive to the decision-making movement of Congress. On the other hand, it demonstrates how politics is creating a routine towards the emotional tone of networks, influenced by the chaotic flow of public opinion on the Internet. 5. ACKNOWLEDGMENTS Funding for the project ‘Mapping Controversies on the Internet: a scientific cooperation between researchers who analyze the Figure 5. Network of profiles that participated in the debate relationship between Aesthetics, Power and Internet’ generously on the Marco Civil from July to December 2013. In the supplied by National Council for Scientific and Technological Spotlight, profiles whose messages were most popular in the Development (CNPq), National Academic Cooperation Program network. (Procad), Coordination of Improvement of Higher Education The graph in Figure (5) shows the relationship established Personnel (Capes), Foundation of the Ministry of Education through retweets from profiles that between July 17 and (MEC). Our thanks to the team at the Laboratory of Studies in December 31 that used the keyword "Marco Civil". To produce Image and Cyberculture (LABIC) for the continues support. this visualization we processed the data with the high gravity scale to bring closer together those actors who had more connections 6. REFERENCES with the group to which they belong. After this first step, we [1] Latour, B. 2005. Reassembling the Social: an Introduction to generated a statistic of modularity in order to visually emphasize Actor Network Theory. Oxford University Press. Oxford. each perspective by assigning each a different color. We used the [2] Venturini, T. 2010. Diving in magma: how to explore metric of authority to give prominence to nodes that had both controversies with actor-network theory. In Public stronger and larger quantities of connections in the network, with Understanding of Science 19 (2009): 1-16. the goal of finding those individuals who had a higher indegree in the Marco Civil controversy. All told, the final goal was to display [3] Passos, N. 2014. O Black Bloc e o papel das mídias sociais those who received the highest number of RTs of other important nas manifestações brasileiras de 7 de setembro de 2013. actors in the network. For these groups, sharing messages creates Unpublished.