International Journal Publications
- Search Bias Quantification: Investigating Political Bias in Social Media and Web Search
Abstract: Users frequently use search systems on the Web as well as online social media to learn about
ongoing events and public opinion on personalities. Prior studies have shown that the
top-ranked results returned by these search engines can shape user opinion about the topic
(e.g., event or person) being searched. In case of polarizing topics like politics,
where multiple competing perspectives exist, the political bias in the top search results
can play a significant role in shaping public opinion towards (or away from) certain perspectives.
Given the considerable impact that search bias can have on the user, we propose a generalizable
search bias quantification framework that not only measures the political bias in ranked list
output by the search system but also decouples the bias introduced by the different
sources — input data and ranking system. We apply our framework to study the political
bias in searches related to 2016 US Presidential primaries in Twitter social media search
and find that both input data and ranking system matter in determining the final search
output bias seen by the users. And finally, we use the framework to compare the relative bias
for two popular search systems — Twitter social media search and Google web search — for queries
related to politicians and political events. We end by discussing some potential solutions to
signal the bias in the search results to make the users more aware of them.
- Managing Longitudinal Exposure of Socially Shared Data on the Twitter Social Media
Abstract: On most online social media sites today, user-generated data remains accessible to allowed
viewers unless and until the data owner changes her privacy preferences. In this paper, we present a large-scale
measurement study focused on understanding how users control the longitudinal exposure of their publicly
shared data on social media sites. Our study, using data from Twitter, finds that a significant fraction
of users withdraw a surprisingly large percentage of old publicly shared data—more than 28% of 6-year old
public posts (tweets) on Twitter are not accessible today. The inaccessible tweets are either selectively
deleted by users or withdrawn by users when they delete or make their accounts private. We also found a
significant problem with the current exposure control mechanisms—even when a user deletes her tweets or
her account, the current mechanisms leave traces of residual activity, i.e., tweets from other users sent
as replies to those deleted tweets or accounts still remain accessible. We show that using this residual
information one can recover significant information about the deleted tweets or even characteristics of
the deleted accounts. To the best of our knowledge, we are the first to study the information leakage resulting
from residual activities of deleted tweets and accounts. Finally, we propose two exposure control mechanisms
that eliminates information leakage via residual activities. One of our mechanisms optimize for allowing
meaningful social interactions with user posts and another mechanism aims to control longitudinal exposure
via anonymization . We discuss the merits and drawbacks of our proposed mechanisms compared to existing
- An Evaluation of Sentiment Analysis for Mobile Devices
Abstract: Sentiment Analysis has become a key tool to extract knowledge from data containing opinions
and sentiments, particularly, data from online social systems. With the increasing use of smartphones to
access social media platforms, a new wave of applications that explore sentiment analysis in the mobile
environment is beginning to emerge. However, there are various existing sentiment analysis methods and
it is unclear which of them are deployable in the mobile environment. In this paper, we provide the first
of a kind study in which we compare the performance of 14 sentence-level sentiment analysis methods in
the mobile environment. To do that, we adapted these methods to run on Android OS and then we measure their
performance in terms of memory, CPU, and battery consumption. Our findings unveil methods that require
almost no adaptations and run relatively fast as well as methods that could not be deployed due to excessive
use of memory. We hope our effort provides a guide to developers and researchers interested in exploring
sentiment analysis as part of a mobile application and can help new applications to be executed without
the dependency of a server-side API. We also share the Android API that implements all the 14 sentiment
analysis used in this paper.
- Longitudinal Privacy Management in Social Media: The Need for Better Controls
Abstract: This large-scale measurement study of Twitter focuses on understanding how users control
the longitudinal exposure of their publicly shared social data — that is, their tweets — and the limitations
of currently used control mechanisms. Our study finds that, while Twitter users widely employ longitudinal
exposure control mechanisms, they face two fundamental problems. First, even when users delete their data
or account, the current mechanisms leave signficant traces of residual activity. Second, these mechanisms
single out withdrawn tweets or accounts, attracting undesirable attention to them. To address both problems,
an inactivity- based withdrawal scheme for improved longitudinal exposure control is explored.
- You followed my bot! Transforming robots into influential users in Twitter
Abstract: Systems like Klout and Twitalyzer were developed as an attempt to measure the influence of
users within social networks. Although the algorithms used by these systems are not public known, they
have been widely used to rank users according to their influence, especially in the Twitter social network.
As media companies might base their viral marketing campaigns on influence scores, users might attempt
to boost their influence scores with simple mechanisms like following unknown users to be followed back
or even interacting with those who reciprocate these actions. In this paper, we investigate if widely used
influence scores are vulnerable and easy to manipulate. Our approach consists of developing Twitter bot
accounts able to interact with real users to verify strategies that can increase their influence scores
according to different systems. Our results show that it is possible to become influential using very simple
strategies, suggesting that these systems should review their influence score algorithms to avoid accounting
with automatic activity.
- (Mis)Information Dissemination in WhatsApp: Gathering, Analyzing and Countermeasures
Abstract: WhatsApp has revolutionized the way people communicate and interact. It is not only cheaper than the traditional Short Message Service (SMS) communication but it also brings a new form of mobile communication: the group chats. Such groups are great forums for collective discussions on a variety of topics. In particular, in events of great social mobilization, such as strikes and electoral campaigns, WhatsApp group chats are very attractive as they facilitate information exchange among interested people. Yet, recent events have raised concerns about the spreading of misinformation in WhatsApp. In this work, we analyze information dissemination within WhatsApp, focusing on publicly accessible political-oriented groups, collecting all shared messages during major social events in Brazil: a national truck drivers' strike and the Brazilian presidential campaign. We analyze the types of content shared within such groups as well as the network structures that emerge from user interactions within and cross-groups. We then deepen our analysis by identifying the presence of misinformation among the shared images using labels provided by journalists and by a proposed automatic procedure based on Google searches. We identify the most important sources of the fake images and analyze how they propagate across WhatsApp groups and from/to other Web platforms.
- WhatsApp Monitor: A Fact-Checking System for WhatsApp
WhatsApp is the most popular communication application in many developing countries such as Brazil, India, and Mexico, where many people use it as an interface to the web. Due to its encrypted and peer-to-peer nature feature, it is hard for researchers to study which content people share through WhatsApp at scale. In this demo paper, we propose WhatsApp Monitor
), a web-based system that helps researchers and journalists explore the nature of content shared on WhatsApp public groups from two different contexts: Brazil and India. Our tool monitors multiple content categories such as images, videos, audio, and textual messages posted on a set of WhatsApp groups and displays the most shared content per day. Our tool has been used for monitoring content during the 2018 Brazilian general election and was one of the major sources for estimating the spread of misinformation and helping fact-checking efforts.
- On Microtargeting Socially Divisive Ads: A Case Study of Russia-Linked Ad Campaigns on Facebook
Filipe N. Ribeiro, Koustuv Saha, Mahmoudreza Babaei, Lucas Henrique, Johnnatan Messias, Fabrício Benevenuto, Oana Goga, Krishna P. Gummadi, and Elissa M. Redmiles.
- In Proceedings of the Conference on Fairness, Accountability, and Transparency (FAT*'19), Atlanta, Georgia.
Abstract: Targeted advertising is meant to improve the efficiency of matching advertisers
to their customers. However, targeted advertising can also be abused by malicious advertisers to
efficiently reach people susceptible to false stories, stoke grievances, and incite social conflict.
Since targeted ads are not seen by non-targeted and non-vulnerable people, malicious ads are likely
to go unreported and their effects undetected. This work examines a specific case of malicious advertising,
exploring the extent to which political ads from the Russian Intelligence Research Agency (IRA) run prior to
2016 U.S. elections exploited Facebook's targeted advertising infrastructure to efficiently target ads on divisive
or polarizing topics (e.g., immigration, race-based policing) at vulnerable sub-populations. In particular,
we do the following: (a) We conduct U.S. census-representative surveys to characterize how users with different
political ideologies report, approve, and
perceive truth in the content of the IRA ads.
Our surveys show that many ads are "divisive": they elicit very different reactions from people belonging to
different socially salient groups. (b) We characterize how these divisive ads are targeted to sub-populations that
feel particularly aggrieved by the status quo. Our findings support existing calls for greater transparency of
content and targeting of political ads. (c) We particularly focus on how the Facebook ad API facilitates such targeting.
We show how the enormous amount of personal data Facebook aggregates about users and makes available to advertisers
enables such malicious targeting.
- White, Man, and Highly Followed: Gender and Race Inequalities in Twitter
Abstract: Social media is considered a democratic space in which people connect and interact with each
other regardless of their gender, race, or any other demographic factor. Despite numerous efforts that
explore demographic factors in social media, it is still unclear whether social media perpetuates old inequalities
from the offline world. In this paper, we attempt to identify gender and race of Twitter users located
in U.S. using advanced image processing algorithms from Face++. Then, we investigate how different demographic
groups (i.e. male/female, Asian/Black/White) connect with other. We quantify to what extent one group follow
and interact with each other and the extent to which these connections and interactions reflect in inequalities
in Twitter. Our analysis shows that users identified as White and male tend to attain higher positions
in Twitter, in terms of the number of followers and number of times in user's lists. We hope our effort
can stimulate the development of new theories of demographic information in the online space.
- Demographics of News Sharing in the U.S. Twittersphere
Abstract: The widespread adoption and dissemination of online news through social media systems have
been revolutionizing many segments of our society and ultimately our daily lives. In these systems, users
can play a central role as they share content to their friends. Despite that, little is known about news
spreaders in social media. In this paper, we provide the first of its kind in-depth characterization of
news spreaders in social media. In particular, we investigate their demographics, what kind of content
they share, and the audience they reach. Among our main findings, we show that males and white users tend
to be more active in terms of sharing news, biasing the news audience to the interests of these demographic
groups. Our results also quantify differences in interests of news sharing across demographics, which has
implications for personalized news digests.
- Linguistic Diversities of Demographic Groups in Twitter
Abstract: The massive popularity of online social media provides a unique opportunity for researchers
to study the linguistic characteristics and patterns of user's interactions. In this paper, we provide
an in-depth characterization of language usage across demographic groups in Twitter. In particular, we
extract the gender and race of Twitter users located in the U.S. using advanced image processing algorithms
from Face++. Then, we investigate how demographic groups (i.e. male/female, Asian/Black/White) differ in
terms of linguistic styles and also their interests. We extract linguistic features from 6 categories (affective
attributes, cognitive attributes, lexical density and awareness, temporal references, social and personal
concerns, and interpersonal focus), in order to identify the similarities and differences in particular
writing set of attributes. In addition, we extract the absolute ranking difference of top phrases between
demographic groups. As a dimension of diversity, we also use the topics of interest that we retrieve from
each user. Our analysis unveils clear differences in the writing styles (and the topics of interest) of
different demographic groups, with variation seen across both gender and race lines. We hope our effort
can stimulate the development of new studies related to demographic information in the online space.
- Who Makes Trends? Understanding Demographic Biases in Crowdsourced Recommendations
Users of social media sites like Facebook and Twitter rely on crowdsourced content recommendation systems (e.g., Trending
Topics) to retrieve important and useful information. Contents selected for recommendation indirectly give
the initial users who promoted (by liking or posting) the content an opportunity to propagate their messages
to a wider audience. Hence, it is important to understand the demographics of people who make a content
worthy of recommendation, and explore whether they are representative of the media site's overall population.
In this work, using extensive data collected from Twitter, we make the first attempt to quantify and explore
the demographic biases in the crowdsourced recommendations. Our analysis, focusing on the selection of
trending topics, finds that a large fraction of trends are promoted by crowds whose demographics are significantly
different from the overall Twitter population. More worryingly, we find that certain demographic groups
are systematically under-represented among the promoters of the trending topics. To make the demographic
biases in Twitter trends more transparent, we developed and deployed a Web-based service "Who-Makes-Trends"
- Quantifying Search Bias: Investigating Sources of Bias for Political Searches in Social Media
Abstract: To help their users to discover the most interesting contents at a particular time, social
media sites like Facebook and Twitter deploy content recommendation systems (such as Trending Topics),
which often rely on crowdsourced popularity signals to select the contents. Once the contents are selected
for recommendation, they reach a large population, effectively giving the initial users of the contents
an opportunity to propagate their messages to the wider public. Hence, it is extremely important to understand
the demographics of people who make a content worthy of recommendation, and explore whether there are demographic
biases in the recommended contents where the majority of the recommended contents were initially popular
with crowds exhibiting skewed demographic distributions.
In this work, using extensive data collected from Twitter, we make the first attempt to quantify and explore
the demographic biases in the crowdsourced recommendations (particularly, in the selection of trending
topics). In our analysis, we find that very different topics are popular among different demographic groups,
and in practice, there is a bias towards a particular demographic while selecting the trending topics.
We further propose and evaluate different techniques to limit such demographic biases in trending topic
- From Migration Corridors to Clusters: The Value of Google+ Data for Migration Studies
Abstract: Recently, there have been considerable efforts to use online data to investigate international
migration. These efforts show that Web data are valuable for estimating migration rates and are relatively
easy to obtain. However, existing studies have only investigated flows of people along migration corridors,
i.e. between pairs of countries. In our work, we use data about "places lived" from millions of Google+
users in order to study migration "clusters", i.e. groups of countries in which individuals have lived.
For the first time, we consider information about more than two countries people have lived in. We argue
that these data are very valuable because this type of information is not available in traditional demographic
sources which record country-to-country migration flows independent of each other. We show that migration
clusters of country triads cannot be identified using information about bilateral flows alone. To demonstrate
the additional insights that can be gained by using data about migration clusters, we first develop a model
that tries to predict the prevalence of a given triad using only data about its constituent pairs. We then
inspect the groups of three countries which are more or less prominent, compared to what we would expect
based on bilateral flows alone. Next, we identify a set of features such as a shared language or colonial
ties that explain which triple of country pairs are more or less likely to be clustered when looking at
country triples. Then we select and contrast a few cases of clusters that provide some qualitative information
about what our data set shows. The type of data that we use is potentially available for a number of social
media services. We hope that this first study about migration clusters will stimulate the use of Web data
for the development of new theories of international migration that could not be tested appropriately before.
- Towards Sentiment Analysis for Mobile Devices
Abstract: The increasing use of smartphones to access social media platforms opens a new wave of applications
that explore sentiment analysis in the mobile environment. However, there are various existing sentiment
analysis methods and it is unclear which of them are deployable in the mobile environment. This paper provides
the first of a kind study in which we compare the performance of 17 sentence-level sentiment analysis methods
in the mobile environment. To do that, we adapted these sentence-level methods to run on Android OS and
then we measure their performance in terms of memory usage, CPU usage, and battery consumption. Our findings
unveil sentence-level methods that require almost no adaptations and run relatively fast as well as methods
that could not be deployed due to excessive use of memory. We hope our effort provides a guide to developers
and researchers interested in exploring sentiment analysis as part of a mobile application and can help
new applications to be executed without the dependency of a server-side API.
- Forgetting in Social Media: Understanding and Controlling Longitudinal Exposure of Socially Shared Data
- A System for Monitoring Public Political Groups in WhatsApp
Abstract: In Brazil, 48% of the population use WhatsApp to share and discuss news.
Currently, there are serious concerns that this platform can become a fertile ground
for groups interested in disseminating misinformation, especially as part of articulated
political campaigns. Particularly, WhatsApp provides an important space for users to engage
in public conversations that worth attention, the public groups. These groups are suitable
for political activism and social movement organization. Additionally, it is reasonable to
assume that a malicious misinformation campaign might attempt to maximize the audience of a
fake story by sharing it in existing public groups. In this paper, we present a system for gathering,
analyzing and visualize public groups in WhatsApp. In addition to describe our methodology,
we also provide a brief characterization of the content shared in 127 Brazilian groups.
We hope our system can help journalists and researchers to understand the repercussion of
events related to the Brazilian elections within these groups.
- Algoritmos de Aprendizado de Máquina para Predição de Resultados das Lutas de MMA
Abstract: This paper proposes using machine learning algorithms to predict the outcome of an MMA fight
based on the characteristics of the two fighters and their recent opponents. Our experimental evaluation
shows an approach to create a dataset applicable to individual sports and one of the evaluated algorithms
has 67% of successful predictions.
- Brazil Around the World: Characterizing and Detecting Brazilian Emigrants Using Google+
Abstract: Currently available data about people whose left their home country to live in a foreign
country does not adequately capture the standards of contemporary global migration flows. A new trend for
migration studies is to study the data from the Internet, either by Social Networks or other data in the
WEB. In this study, we collected users data from the social network Google+ to investigate which features
of Brazilian users are relevant to classify them as a possible emigrant. Our study uses machine learning
techniques, SVM. We selected some features to compose our dataset. Our results show that the network features
were the ones that had greater capacity for discrimination. The most relevant for the prediction of Brazilian
emigrants users are, in order: reciprocity, PageRank, in-degree, clustering coefficient and ratio of incoming
- Bazinga! Caracterizando e Detectando Sarcasmo e Ironia no Twitter
Pollyanna Gonçalves, Daniel Dalip, Julio C. S. Reis, Johnnatan Messias, Filipe N. Ribeiro, Philipe Melo, Leandro A. A. Silva, Marcos Gonçalves, and Fabrício Benevenuto.
- In Proceedings of the Proceedings of the Brazilian Workshop on Social Network Analysis and Mining (BraSNAM).
Recife, Brazil. July, 2015.
Abstract: Sarcasm and irony are widely used forms of speech used inside and outside the Web, having
the power to transform a sentence regarding its polarity or sense. The ability of characterizing and detecting
sarcastic and ironic messages on data collected from Web could improve many decision-making systems based
on Natural Language Processing (NLP) such as the sentiment analysis, text summarization and review ranking
systems. In this work, we propose some approaches to the task of characterization and detection of sarcasm
and irony in messages posted on Twitter online social network. Using an automatically collected dataset
with the hashtags “#sarcasm” and“#irony”, and by exploiting a large set of characterization and classification
techniques, our results show satisfactory rates of accuracy and Macro-F1.
- Bots Sociais: Como robôs podem se tornar pessoas influentes no Twitter?
- Sigam-me os bons! Transformando robôs em pessoas influentes no Twitter
Abstract: Systems that classify influential users in social networks has been used with great frequency,
being referenced in scientific papers and the media as the ideal standard for evaluation of influence in
the social network Twitter. We consider this measure a complex and subjective and therefore suspect vulnerability
and ease of handling these systems. Based on this, we performed experiments and analyzes in two ranking
systems of influence: Klout and Twitalyzer. We create simple robots capable of interacting through Twitter
accounts and measure their influence. Our results show that it is possible to be influential through simple
strategies. This suggests that the systems do not have ideal metric to rank influence.
- Characterizing Interconnections and Linguistic Patterns in Twitter
Abstract: Social media is considered a democratic space in which people connect and interact with each
other regardless of their gender, race, or any other demographic aspect. Despite numerous efforts that
explore demographic aspects in social media, it is still unclear whether social media perpetuates old inequalities
from the offline world. In this dissertation, we attempt to identify gender and race of Twitter users located
in the United States using advanced image processing algorithms from Face++. We investigate how different
demographic groups (i.e. male/female, asian/black/white) connect with each other and differentiate them
regarding linguistic styles and also their interests. We quantify to what extent one group follows and
interacts with each other and the extent to which these connections and interactions reflect in inequalities
in Twitter. We also extract linguistic features from six categories (affective attributes, cognitive attributes,
lexical density and awareness, temporal references, social and personal concerns, and interpersonal focus)
in order to identify the similarities and the differences in the messages they share in Twitter. Furthermore,
we extract the absolute ranking difference of top phrases between demographic groups. As a dimension of
diversity, we also use the topics of interest that we retrieve from each user. Our analysis shows that
users identified as white and male tend to attain higher positions, in terms of the number of followers
and number of times in another user's lists, in Twitter. There are clear differences in the way of writing
across different demographic groups in both gender and race domains as well as in the topic of interest.
We hope our effort can stimulate the development of new theories of demographic information in the online
space. Finally, we developed a Web-based system that leverages the demographic aspects of users to provide
transparency to the Twitter trending topics system.
- Framework Para Sistemas de Navegação de Veículos Aéreos Não Tripulados
Abstract: Become autonomous unmanned flights undoubtedly enable new opportunities for scientific development.
The drones can be used in military services, for example, in combat or as well as for rescue missions,
aerial survey, supervision and inspection of a territory, attracting significant attention from media outlets
such as, for example, television stations, radio, newspapers and internet. The goal of this project is
whether it is possible to make viable autonomous flights at AR.Drone 2.0 and the understanding of its operation.
This will require the implementation of a control program for autonomous flights. This framework requires
the acquisition of data during the flight, which are obtained using sensors which use Arduino. The Arduino
communication with the drone is needed for the inclusion of new sensors and the use of the AR.Drone is
performed by the framework Node.js. Each remote button has a specific command, and may be in order for
the user to create own missions or even perform some missions previously implemented by the developer.
All tests were run on the AR.Drone 2.0, using the Node.js framework, sensors and a remote control. Through
the experiments and presented studies became possible to achieve the proposed objective, making possible
the implementation of autonomous flights in drone. As a result, for the realization of autonomous flight
we designed a framework where the user can create autonomous flight missions for the drone run them. These
commands are sent to the drone by the user due to use of a remote control. This remote control sends data
to a sensor connected to the Arduino that processes the data and then is read and interpreted by the drone.
For more information and complete curriculum visit: