Project

Social Networks

Updates
0 new
0
Recommendations
0 new
0
Followers
0 new
10
Reads
0 new
143

Project log

Ponnurangam Kumaraguru
added 2 research items
On 6 January 2021, a mob of right-wing conservatives stormed the USA Capitol Hill interrupting the session of congress certifying 2020 Presidential election results. Immediately after the start of the event, posts related to the riots started to trend on social media. A social media platform which stood out was a free speech endorsing social media platform Parler; it is being claimed as the platform on which the riots were planned and talked about. Our report presents a contrast between the trending content on Parler and Twitter around the time of riots. We collected data from both platforms based on the trending hashtags and draw comparisons based on what are the topics being talked about, who are the people active on the platforms and how organic is the content generated on the two platforms. While the content trending on Twitter had strong resentments towards the event and called for action against rioters and inciters, Parler content had a strong conservative narrative echoing the ideas of voter fraud similar to the attacking mob. We also find a disproportionately high manipulation of traffic on Parler when compared to Twitter.
Social media has grown exponentially in a short period, coming to the forefront of communications and online interactions. Despite their rapid growth, social media platforms have been unable to scale to different languages globally and remain inaccessible to many. In this report, we characterize Koo, a multilingual micro-blogging site that rose in popularity in 2021, as an Indian alternative to Twitter. We collected a dataset of 4.07 million users, 163.12 million follower-following relationships, and their content and activity across 12 languages. The prominent presence of Indian languages in the discourse on Koo indicates the platform's success in promoting regional languages. We observe Koo's follower-following network to be much denser than Twitter's, comprising of closely-knit linguistic communities. This initial characterization heralds a deeper study of the dynamics of the multilingual social network and its diverse Indian user base.
Rishabh Kaushal
added a research item
Users have their accounts on multiple Online Social Networks (OSNs) to access a variety of content and connect to their friends. Consequently, user behaviors get distributed across many OSNs. Collection of comprehensive user information referred to as user profiling; an essential first step is to link user accounts (identities) belonging to the same individual across OSNs. To this end, we provide a detailed methodology of five methods useful for user profiling, which we refer to as Advanced Search Operator (ASO), Social Aggregator (SA), Cross-Platform Sharing (CPS), Self-Disclosure (SD) and Friend Finding Feature (FFF). Taken together, we collect linked identities of 208,120 individuals distributed across 43 different OSNs. We compare these methods quantitatively based on social network coverage and the number of linked identities obtained per-individual. And also perform a qualitative assessment of linked user data, thus obtained by these methods, on the criteria of completeness, validity, consistency, accuracy, and timeliness.
Rishabh Kaushal
added a research item
Users create accounts on multiple social networks to get connected to their friends across these networks. We refer to these user accounts as user identities. Since users join multiple social networks, therefore, there will be cases where a pair of user identities across two different social networks belong to the same individual. We refer to such pairs as Cross-Network Linkages (CNLs). In this work, we model the social network as a graph to explore the question, whether we can obtain effective social network graph representation such that node embeddings of users belonging to CNLs are closer in embedding space than other nodes, using only the network information. To this end, we propose a modular and flexible node embedding framework, referred to as NeXLink, which comprises of three steps. First, we obtain local node embeddings by preserving the local structure of nodes within the same social network. Second, we learn the global node embeddings by preserving the global structure, which is present in the form of common friendship exhibited by nodes involved in CNLs across social networks. Third, we combine the local and global node embeddings, which preserve local and global structures to facilitate the detection of CNLs across social networks. We evaluate our proposed framework on an augmented (synthetically generated) dataset of 63,713 nodes & 817,090 edges and real-world dataset of 3338 Twitter-Foursquare node pairs. Our approach achieves an average Hit@1 rate of 98% for detecting CNLs across social networks and significantly outperforms previous state-of-the-art methods.
Ponnurangam Kumaraguru
added a research item
Instagram is a significant platform for users to share media; reflecting their interests. It is used by marketers and brands to reach their potential audience for advertisement. The number of likes on posts serves as a proxy for social reputation of the users, and in some cases, social media influencers with an extensive reach are compensated by marketers to promote products. This emerging market has led to users artificially bolstering the likes they get to project an inflated social worth. In this study, we enumerate the potential factors which contribute towards a genuine like on Instagram. Based on our analysis of liking behaviour, we build an automated mechanism to detect fake likes on Instagram which achieves a high precision of 83.5%. Our work serves an important first step in reducing the effect of fake likes on Instagram influencer market.
Ponnurangam Kumaraguru
added a research item
The social connections people form online affect the quality of information they receive and their online experience. Although a host of socioeconomic and cognitive factors were implicated in the formation of offline social ties, few of them have been empirically validated, particularly in an on-line setting. In this study, we analyze a large corpus of geo-referenced messages, or tweets, posted by social media users from a major US metropolitan area. We linked these tweets to US Census data through their locations. This allowed us to measure emotions expressed in the tweets posted from an area, the structure of social connections, and also use that area's socioeconomic characteristics in analysis. We find that at an aggregate level, places where social media users engage more deeply with less diverse social contacts are those where they express more negative emotions, like sadness and anger. Demographics also has an impact: these places have residents with lower household income and education levels. Conversely, places where people engage less frequently but with diverse contacts have happier, more positive messages posted from them and also have better educated, younger, more affluent residents. Results suggest that cognitive factors and offline characteristics affect the quality of online interactions. Our work highlights the value of linking social media data to traditional data sources, such as US Census, to drive novel analysis of online behavior.
Rishabh Kaushal
added 2 research items
Today, the use of Online Social Media (OSM) is not restricted to merely networking and socializing. Recent events all around the globe attest to the prevalence of use of OSM sites for bringing about dramatic and drastic reforms in real world, phenomenon being referred as Cyber Activism. The real world is marred with various turmoils and people hold myriad variety of views and judgments regarding various issues. Their opinions are often poorly backed by facts. We refer to such inconclusive judgements that users generate and propagate on OSM platforms as Multi-Opinionated Content. One of the greatest challenge in such an environment is to analyze and classify such content into multiple opinion classes. In this work, we propose a generic semi-supervised classification based methodology for analyzing and classifying multi-opinionated content. We have used widely known off-the-shelf classifiers namely KNN, decision tree and random forest in our approach. To implement and validate our methodology, we have mined opinions on content in various forms, namely videos, tweets and posts on three popular social media platforms namely Youtube, Twitter and Facebook, respectively. In our validation, we have taken the Kashmir conflict between India and Pakistan as our case study. We have used plethora of features in building the classification model. Our experiments show that Random Forest classifier gives maximum accuracy of 90.02% and user level features give the best results. Our work can be used to process large amount of multi-opinionated content for effective and accurate decision making in the era of cyber activism generating multi-opinionated content.
Ponnurangam Kumaraguru
added a research item
Online social media has become an integral part of every Internet users' life. It has given common people a platform and forum to share information, post their opinions and promote campaigns. The threat of exploitation of social media like Facebook, Twitter, etc. by malicious entities, becomes crucial during a crisis situation, like bomb blasts or natural calamities such as earthquakes and floods. In this report, we attempt to characterize and extract patterns of activity of general users on Twitter during a crisis situation. This is the first attempt to study an India-centric crisis event such as the triple bomb blasts in Mumbai (India), using online social media. In this research, we perform content and activity analysis of content posted on Twitter after the bomb blasts. Through our analysis, we conclude, that the number of URLs and @-mentions in tweets increase during the time of the crisis in comparison to what researchers have exhibited for normal circumstances. In addition to the above, we empirically show that the number of tweets or updates by authority users (those with large number of followers) are very less, i.e. majority of content generated on Twitter during the crisis comes from non authority users. In the end, we discuss certain case scenarios during the Mumbai blasts, where rumors were spread through the network of Twitter.
Rishabh Kaushal
added a research item
The last decade has witnessed a boom in social networking platforms; each new platform is unique in its own ways, and offers a different set of features and services. In order to avail these services, users end up creating multiple virtual identities across these platforms. Researchers have proposed numerous techniques to resolve multiple such identities of a user across different platforms. However, the ability to link different identities poses a threat to the users’ privacy; users may or may not want their identities to be linkable across networks. In this paper, we propose Nudging Nemo, a framework which assists users to control the linkability of their identities across multiple platforms. We model the notion of linkability as the probability of an adversary (who is part of the user’s network) being able to link two profiles across different platforms, to the same real user. Nudging Nemo has two components; a linkability calculator which uses state-of-the-art identity resolution techniques to compute a normalized linkability measure for each pair of social network platforms used by a user, and a soft paternalistic nudge, which alerts the user if any of their activity violates their preferred linkability. We evaluate the effectiveness of the nudge by conducting a controlled user study on privacy conscious users who maintain their accounts on Facebook, Twitter, and Instagram. Outcomes of user study confirmed that the proposed framework helped most of the participants to take informed decisions, thereby preventing inadvertent exposure of their personal information across social network services.
Rishabh Kaushal
added 3 research items
Existing travel related systems and commonly used websites have some major limitations which cause efforts to be made by the traveler before going out on vacation. Some of these sites allow users to write their personal experiences about visited places but don’t produce a proper itinerary, and those which do, focus only on minimizing the travel time between POIs ignoring other important factors like POI ratings, traffic conditions, etc. Our work focuses on Building Real-Time Travel Itineraries using ‘off-the-shelf’ data from the Web. The proposed solution solves the existing limitations by using an optimization algorithm, which produces a real-time itinerary after optimizing various important factors like travel time between POIs, traffic conditions, ratings of POIs, to enhance the traveler’s experience in a city. Out of the several optimization approaches available, an algorithm was finalized after comparison of performance and accuracy between the approaches. Best results were obtained in case of a dynamic programming based approach, which optimized both accuracy and performance.
In recent times, Online Social Networks (OSNs) have gained immense popularity among users of all age groups, particularly the young. Due to their widespread popularity and engagement, we are observing a rapid rise in the number of OSNs and the wide variety of services (or ecosystem) that they offer. However, at the same time, there are growing concerns about availability, visibility and leakage of Personally Identifiable Information (PII) of users on these OSNs. In this paper, we focus our attention on the OSNs which offer dating services. In dating OSNs, users typically don’t reveal their true identities in their profiles. In our work, we first propose a framework to extract user’s PII through targeted intelligent conversations. Subsequently, we present detailed evaluation of vulnerability of users to PII disclosures. Our field study on 100 users validates the claim that these users are extremely vulnerable to PII disclosures.
Rishabh Kaushal
added 9 research items
Online Social Networks (OSNs) are deemed to be the most sought-after societal tool used by the masses world over to communicate and transmit information. Our dependence on these platforms for seeking opinions, news, updates, etc. is increasing. While it is true that OSNs have become a new medium for dissemination of information, at the same time, they are also fast becoming a playground for the spread of misinformation, propaganda, fake news, rumors, unsolicited messages, etc. Consequently, we can say that an OSN platform comprises of two kinds of users namely, Spammers and Non-Spammers. Spammers, out of malicious intent, post either unwanted (or irrelevant) information or spread misinformation on OSN platforms. As part of our work, we propose mechanisms to detect such users (Spammers) in Twitter social network (a popular OSN). Our work is based on a number of features at tweet-level and user-level like Followers/Followees, URLs, Spam Words, Replies and HashTags. In our work, we have applied three learning algorithms namely Naive Bayes, Clustering and Decision trees. Furthermore, to improve detection of Spammers, a novel integrated approach is proposed which 'combines' the advantages of the three learning algorithms mentioned above. Improvement of spam detection is measured on the basis of Total Accuracy, Spammers Detection Accuracy and Non-Spammers Detection Accuracy. Results, thus obtained, show that our novel integrated approach that combines all algorithms outperforms other classical approaches in terms of overall accuracy and detect Non-Spammers with 99% accuracy with an overall accuracy of 87.9%.