Social Media Streams Research Articles

A survey on cost-effective context-aware distribution of social data streams over energy-efficient data centres

Social media have emerged in the last decade as a viable and ubiquitous means of communication. The ease of user content generation within these platforms, e.g. check-in information, multimedia data, etc., along with the proliferation of Global Positioning System (GPS)-enabled, always-connected capture devices lead to data streams of unprecedented amount and a radical change in information sharing. Social data streams raise a variety of practical challenges, including derivation of real-time meaningful insights from effectively gathered social information, as well as a paradigm shift for content distribution with the leverage of contextual data associated with user preferences, geographical characteristics and devices in general. In this article we present a comprehensive survey that outlines the state-of-the-art situation and organizes challenges concerning social media streams and the infrastructure of the data centres supporting the efficient access to data streams in terms of content distribution, data diffusion, data replication, energy efficiency and network infrastructure. We systematize the existing literature and proceed to identify and analyse the main research points and industrial efforts in the area as far as modelling, simulation and performance evaluation are concerned.

The Media and the 2016 Election (Presentation Slides)

The origins of this paper emanated from an interdisciplinary conference at the University of Houston that accompanied the 2016 Republican Presidential Debate that was held on campus. The paper was then revised for a presentation in Mexico City at the Matias Romero Institute in the summer of 2017. On August 6, 2018, the paper was presented at the SEALS Conference in Fort Lauderdale, Florida. Attached are slides from that presentation. This author is very appreciative of the positive feedback received during these presentations. In examining this issue, the paper outlines the power imbalance that existed between the traditional media and the presidential candidates during the Republican primaries as well as during the general election. This power imbalance provided Candidate Trump with an unprecedented amount of “earned” media coverage. This was partly due to his celebrity status, which the traditional media commodified through its coverage fueling their increased advertising rates. During the 2016 election, the traditional media experienced unprecedented viewership during the Republican debates and news coverage of Candidate Trump. Candidate Trump’s celebrity status and nontraditional approach to the election campaign fueled a cycle of the media disproportionately covering Candidate Trump’s campaign in comparison to his Republican rivals in the primaries and Candidate Clinton’s in the general election. Moreover, Candidate Trump skillfully used Twitter and Facebook and used these social media resources more often than his Republican rivals in the Republican primaries and against Candidate Clinton in the general election. This allowed the Trump campaign to earn even more coverage from traditional media allowing him to win the Republican nomination and the general election by raising and spending less money than his respective rivals in those contests. In contrast, in the general election, even though Candidate Clinton raised more campaign money than Candidate Trump, Candidate Trump had the significant advantage in earned media. Moreover, unlike Candidate Trump, Candidate Clinton relied more on spending on traditional media advertising than social media. Social media like Facebook and Twitter experienced outside and foreign influence. Facebook acknowledged that “fake news” was distributed on its network in its feed to subscribers. There were also dark advertising to targeted slivers of the population based on messages that either would make the recipients of the ads less likely to vote or more likely. The most troubling aspect about these dark social media ads is that because they were “dark,” there was no opportunity for a rival campaign to counter the ads’ message. After the 2016 election, Facebook acknowledged the manipulation of its network and instituted some measures to self- regulate. Given that so many more members of the voting public get their news from personalized social media streams, these fundamental changes in the media landscape and their use raise the issue of whether self-regulation by the social media companies is sufficient to provide a level playing field among presidential candidates and to prevent outside foreign influence.

Modeling Popularity in Asynchronous Social Media Streams with Recurrent Neural Networks

Understanding and predicting the popularity of online itemsis an important open problem in social media analysis. Considerable progress has been made recently in data-driven predictions, and in linking popularity to external promotions. However, the existing methods typically focus on a single source of external influence, whereas for many types of online content such as YouTube videos or news articles, attention is driven by multiple heterogeneous sources simultaneously – e.g. microblogs or traditional media coverage. Here, we propose RNN-MAS, a recurrent neural network for modeling asynchronous streams. It is a sequence generator that connects multiple streams of different granularity via joint inference. We show RNN-MAS not only outperforms the current state-of-the-art Youtube popularity prediction system by 17%, but also captures complex dynamics, such as seasonal trends of unseen influence. We define two new metrics: the promotion score quantifies the gain in popularity from one unit of promotion for a Youtube video; the loudness level captures the effects of a particular user tweeting about the video. We use the loudness level to compare the effects of a video being promoted by a single highly-followed user (in the top 1% most followed users) against being promoted by a group of mid-followed users. We find that results depend on the type of content being promoted: superusers are more successful in promoting Howto and Gaming videos, whereas the cohort of regular users are more influential for Activism videos. This work provides more accurate and explainable popularity predictions, as well as computational tools for content producers and marketers to allocate resources for promotion campaigns.

Challenges of event detection from social media streams

The area of Event Detection (ED) has attracted researchers' attention over the last few years because of the wide use of social media. Many studies have examined the problem of ED in various social media platforms, like Twitter, Facebook, YouTube, etc. The ED task for social networks involves many issues, including the processing of huge volumes of data with a high level of noise, data collection and privacy issues, etc. Hence, this article discusses and presents the wide range of challenges encountered in the ED process from unstructured text data for the most popular Social Networks (SNs), such as Facebook and Twitter. The main goal is to aid the researchers to understand the main challenges and to discuss the future directions in the ED area.

Collecting Typhoon Disaster Information from Twitter Based on Query Expansion

Social media is a popular source of volunteered geographic information owing to its massive real-time data; however, the use of social media data in the context of geospatial analysis is challenging because complex semantic filters are required for the aggregation of geographic messages from the data streams. This article proposes a new query expansion method for social media streams which updates the query keywords periodically by the words extracted from the preceding search results. The proposed method has optimized the trade-off between precision and coverage of geographical messages by factoring in the influences of the keyword number and refresh cycle in the query process, and some improvements on the classic Term Frequency-Inverse Document Frequency (TF-IDF) method for short texts were achieved. Furthermore, a number of filters based upon relevance to the target topic were established and tested. This method was tested on a dataset from Twitter within the geographic extent of Macau in August 2017 during two consecutive typhoon hits. The result supports its effectiveness with a controllable precision and considerable increment of relevant information. Moreover, the query keywords can adjust themselves to the local language environment by discovering new keywords. To conclude, this query expansion method is able to provide a reliable method for social media-based information retrieval.

Early detection and information extraction for weather-induced floods using social media streams

Today we are using an unprecedented wealth of social media platforms to generate and share information regarding a wide class of events, which include extreme meteorological conditions and natural hazards such as floods. This paper proposes an automated set of services that start from the availability of weather forecasts, including both an event detection technique and a selective information retrieval from on-line social media. The envisioned services aim to provide qualitative feedback for meteorological models, detect the occurrence of an emergency event and extract informative content that can be used to complement the situational awareness. We implement such services and evaluate them during a recent weather induced flood. Our approach could be highly beneficial for monitoring agencies and meteorological offices, who act in the early warning phase, and also for authorities and first responders, who manage the emergency response phase.

Early detection and information extraction for weather-induced floods using social media streams

Early detection and information extraction for weather-induced floods using social media streams

Sentinel: A Codesigned Platform for Semantic Enrichment of Social Media Streams

We introduce the Sentinel platform that supports semantic enrichment of streamed social media data for the purposes of situational understanding. The platform is the result of a codesign effort between computing and social scientists, iteratively developed through a series of pilot studies. The platform is founded upon a knowledge-based approach, in which input streams ( channels ) are characterized by spatial and terminological parameters, collected media is preprocessed to identify significant terms ( signals ), and data are tagged ( framed ) in relation to an ontology. Interpretation of processed media is framed in terms of the 5W framework (who, what, when, where, and why). The platform is designed to be open to the incorporation of new processing modules, building on the knowledge-based elements (channels, signals, and framing ontology) and accessible via a set of user-facing apps. We present the conceptual architecture for the platform, discuss the design and implementation challenges of the underlying stream-processing system, and present a number of apps developed in the context of the pilot studies, highlighting the strengths and importance of the codesign approach and indicating promising areas for future research.

Semantic Interoperability for Data Analysis in the Food Supply Chain

Food supply chains consist of many links and operate on a global scale with many stakeholders involved from farm to fork. Each stakeholder maintains data about food products that they handle, but this data is not transparently available to all stakeholders in the chain and trust in data sharing is low. In addition, there are various other data sources that contain interesting data for stakeholders in the food chain, such as import/export transactions, production (forecast) data, parcel crop information, local weather predictions and social media streams. To improve their production, growers and traders are very interested in trends in the market and activities in supply and demand. To make all stakeholders in the food chain benefit from these data sources and to share data more transparently, the Dutch horticulture and food domain is developing the HortiCube platform via which various data sources are made accessible to application developers using a secure, linked data application interface. This paper describes the design and engineering of the semantic approach to enable interoperability between data sources. This includes (1) a high-level design of the HortiCube, (2) the metadata ontology used for describing the contents of the data sources in the HortiCube, (3) the common horticulture model used to achieve semantic alignment between data sources in the HortiCube, (4) a test application for a specific product case and (5) a discussion of our results and future work on this topic. The main contribution of our research is the generic solution and ontology design to the semantic challenges that arise when different data sources are combined to answer analysis questions for the user.

Leveraging Localized Social Media Insights for Industry Early Warning Systems

Social Media (SM) has become the easiest, cheapest and fastest channel for companies to identify the events that affect their customers. The geo-location capabilities of the SM interactions enable Early Warning Systems to alert not only when the quality of service decays, but also where and how many customers are impacted. In this paper we present a system and a set of supporting metrics that exploit the geo-localized SM stream, quantify the perceived impact of events, incidents, etc. on a particular area over time. Industrial service providers can add this perceptional perspective to their standard monitoring tools to enable a prompt and appropriate reaction, the decision-making in marketing activities and to unveil customer acquisition opportunities applying the system to the competitors’ customers.

Hazardous Weather Prediction and Communication in the Modern Information Environment

AbstractDuring the last few decades, scientific capabilities for understanding and predicting weather and climate risks have advanced rapidly. At the same time, technological advances, such as the Internet, mobile devices, and social media, are transforming how people exchange and interact with information. In this modern information environment, risk communication, interpretation, and decision-making are rapidly evolving processes that intersect across space, time, and society. Instead of a linear or iterative process in which individual members of the public assess and respond to distinct pieces of weather forecast or warning information, this article conceives of weather prediction, communication, and decision-making as an interconnected dynamic system. In this expanded framework, information and uncertainty evolve in conjunction with people’s risk perceptions, vulnerabilities, and decisions as a hazardous weather threat approaches; these processes are intertwined with evolving social interactions in the physical and digital worlds. Along with the framework, the article presents two interdisciplinary research approaches for advancing the understanding of this complex system and the processes within it: analysis of social media streams and computational natural–human system modeling. Examples from ongoing research are used to demonstrate these approaches and illustrate the types of new insights they can reveal. This expanded perspective together with research approaches, such as those introduced, can help researchers and practitioners understand and improve the creation and communication of information in atmospheric science and other fields.

Study protocol: responding to the needs of patients with IgA nephropathy, a social media approach.

BackgroundIgA nephropathy is the most common cause of glomerulonephritis in the Western world and predominantly affects young adults. Demographically these patients are the biggest users of social media. With increasing numbers of patients turning to social media to seek information and support in dealing with their disease, analysis of social media streams is an attractive modern strategy for understanding and responding to unmet patient need.MethodsTo identify unmet patient need in this population, a framework analysis will be undertaken of prospectively acquired social media posts from patients with IgA nephropathy, acquired from a range of different social media platforms. In collaboration with patients and members of the clinical multidisciplinary team, resources will be created to bridge gaps in patient knowledge and education identified through social media analysis and returned to patients via social media channels and bespoke websites. Analysis of the impact of these resources will be undertaken with further social media analysis, surveys and focus groups.ConclusionsPatients with chronic diseases are increasingly using social networking channels to connect with others with similar diseases and to search for information to help them understand their condition. This project is a 21st century digital solution to understanding patient need and developing resources in partnership with patients, and has wide applicability as a future model for understanding patient needs in a variety of conditions.

A graph-based approach to ememes identification and tracking in Social Media streams

A meme, as defined by Richard Dawkins, is a unit of information, a concept or an idea that spreads from person to person within a culture. Examples of memes can be a musical melody, a catchy phrase, trending news, behavioral patterns, etc. In this article the task of identifying potential memes in a stream of texts is addressed: in particular, the content generated by users of Social Media is considered as a rich source of information offering an updated window on the world happenings and on opinions of people. A textual electronic meme, a.k.a. ememe, is here considered as a frequently replicated set of related words that propagates through the Web over time. In this article an approach is proposed that aims to identify ememes in Social Media streams represented as graph of words. Furthermore, a set of measures is defined to track the change of information in time.

A cloud-enabled automatic disaster analysis system of multi-sourced data streams: An example synthesizing social media, remote sensing and Wikipedia data

Social media streams and remote sensing data have emerged as new sources for tracking disaster events, and assessing their damages. Previous studies focus on a case-by-case approach, where a specific event was first chosen and filtering criteria (e.g., keywords, spatiotemporal information) are manually designed and used to retrieve relevant data for disaster analysis. This paper presents a framework that synthesizes multi-sourced data (e.g., social media, remote sensing, Wikipedia, and Web), spatial data mining and text mining technologies to build an architecturally resilient and elastic solution to support disaster analysis of historical and future events. Within the proposed framework, Wikipedia is used as a primary source of different historical disaster events, which are extracted to build an event database. Such a database characterizes the salient spatiotemporal patterns and characteristics of each type of disaster. Additionally, it can provide basic semantics, such as event name (e.g., Hurricane Sandy) and type (e.g., flooding) and spatiotemporal scopes, which are then tuned by the proposed procedures to extract additional information (e.g., hashtags for searching tweets), to query and retrieve relevant social media and remote sensing data for a specific disaster. Besides historical event analysis and pattern mining, the cloud-based framework can also support real-time event tracking and monitoring by providing on-demand and elastic computing power and storage capabilities. A prototype is implemented and tested with data relative to the 2011 Hurricane Sandy and the 2013 Colorado flooding.

Event summarization for sports games using twitter streams

Given a textual data stream related to an event, social event summarization aims to generate an informative textual description that can capture all the important moments, and it plays a critical role in mining and analyzing social media streams. In this paper, we present a general social event summarization framework using Twitter streams. The proposed framework consists of three key components: participant detection, sub-event detection, and summary tweet extraction. To make the system applicable in real data, an online clustering approach is developed for participant detection and an online temporal-content mixture model is proposed to conduct sub-event detection. Experiments show that the proposed framework can achieve similar performance with its batch counterpart.

Tracking and Analyzing Individual Distress Following Terrorist Attacks Using Social Media Streams.

Risk research has theorized a number of mechanisms that might trigger, prolong, or potentially alleviate individuals' distress following terrorist attacks. These mechanisms are difficult to examine in a single study, however, because the social conditions of terrorist attacks are difficult to simulate in laboratory experiments and appropriate preattack baselines are difficult to establish with surveys. To address this challenge, we propose the use of computational focus groups and a novel analysis framework to analyze a social media stream that archives user history and location. The approach uses time-stamped behavior to quantify an individual's preattack behavior after an attack has occurred, enabling the assessment of time-specific changes in the intensity and duration of an individual's distress, as well as the assessment of individual and social-level covariates. To exemplify the methodology, we collected over 18 million tweets from 15,509 users located in Paris on November 13, 2015, and measured the degree to which they expressed anxiety, anger, and sadness after the attacks. The analysis resulted in findings that would be difficult to observe through other methods, such as that news media exposure had competing, time-dependent effects on anxiety, and that gender dynamics are complicated by baseline behavior. Opportunities for integrating computational focus group analysis with traditional methods are discussed.

Aspect-Based Extraction and Analysis of Affective Knowledge from Social Media Streams

Extracting and analyzing affective knowledge from social media in a structured manner is a challenging task. Decision makers require insights into the public perception of a company's products and services, as a strategic feedback channel to guide communication campaigns, and as an early warning system to quickly react in the case of unforeseen events. The approach presented in this article goes beyond bipolar metrics of sentiment. It combines factual and affective knowledge extracted from rich public knowledge bases to analyze emotions expressed toward specific entities (targets) in social media. The authors obtain common and common-sense domain knowledge from DBpedia and ConceptNet to identify potential sentiment targets. They employ affective knowledge about emotional categories available from SenticNet to assess how those targets and their aspects (such as specific product features) are perceived in social media. An evaluation shows the usefulness and correctness of the extracted domain knowledge, which is used in a proof-of-concept data analytics application to investigate the perception of car brands on social media in the period between September and November 2015.

The xLiMe system: Cross-lingual and cross-modal semantic annotation, search and recommendation over live-TV, news and social media streams

Modern Web search engines still have many limitations: search terms are not disambiguated, search terms in one query cannot be in different languages, the retrieved media items have to be in the same language as the search terms and search results are not integrated across a live stream of different media channels, including TV, online news and social media. The system described in this paper enables all of this by combining a media stream processing architecture with cross-lingual and cross-modal semantic annotation, search and recommendation. All those components were developed in the xLiMe project.

Event Classification in Microblogs via Social Tracking

Social media websites have become important information sharing platforms. The rapid development of social media platforms has led to increasingly large-scale social media data, which has shown remarkable societal and marketing values. There are needs to extract important events in live social media streams. However, microblogs event classification is challenging due to two facts, i.e., the short/conversational nature and the incompatible meanings between the text and the corresponding image in social posts, and the rapidly evolving contents. In this article, we propose to conduct event classification via deep learning and social tracking. First, we introduce a Multi-modal Multi-instance Deep Network (M 2 DN) for microblogs classification, which is able to handle the weakly labeled microblogs data oriented from the incompatible meanings inside microblogs. Besides predicting each microblogs as predefined events, we propose to employ social tracking to extract social-related auxiliary information to enrich the testing samples. We extract a set of candidate-relevant microblogs in a short time window by using social connections, such as related users and geographical locations. All these selected microblogs and the testing data are formulated in a Markov Random Field model. The inference on the Markov Random Field is conducted to update the classification results of the testing microblogs. This method is evaluated on the Brand-Social-Net dataset for classification of 20 events. Experimental results and comparison with the state of the arts show that the proposed method can achieve better performance for the event classification task.

Sub-story detection in Twitter with hierarchical Dirichlet processes

Social media has now become the de facto information source on real world events. The challenge, however, due to the high volume and velocity nature of social media streams, is in how to follow all posts pertaining to a given event over time – a task referred to as story detection. Moreover, there are often several different stories pertaining to a given event, which we refer to as sub-stories and the corresponding task of their automatic detection – as sub-story detection. This paper proposes hierarchical Dirichlet processes (HDP), a probabilistic topic model, as an effective method for automatic sub-story detection. HDP can learn sub-topics associated with sub-stories which enables it to handle subtle variations in sub-stories. It is compared with state-of-the-art story detection approaches based on locality sensitive hashing and spectral clustering. We demonstrate the superior performance of HDP for sub-story detection on real world Twitter data sets using various evaluation measures. The ability of HDP to learn sub-topics helps it to recall the sub-stories with high precision. This has resulted in an improvement of up to 60% in the F-score performance of HDP based sub-story detection approach compared to standard story detection approaches. A similar performance improvement is also seen using an information theoretic evaluation measure proposed for the sub-story detection task. Another contribution of this paper is in demonstrating that considering the conversational structures within the Twitter stream can bring up to 200% improvement in sub-story detection performance.