NSF EAGER: Data-Driven Framework for Analyzing User Interactions in Social Media

社交网络在全球拥有数亿用户，为社交互动、娱乐、学习以及政治和社会变革提供了令人难以置信的机会。因此，人们对了解在线社交网络上的信息传播越来越感兴趣。由于目前许多社交互动都是在在线网络中进行的，因此社会科学家可以获得前所未有的大量有关社交互动的信息。在此类在线网络出现之前，这些调查需要资源密集型活动，例如随机试验、调查和手动数据收集，以收集甚至很小的数据集。现在，记录了大量有关社交网络和社交互动的信息。如此丰富的数据可以让社会科学家以前所未有的规模和详细程度研究社会互动。然而，社会科学家传统上并未接受过处理在线社交网络产生的大量数据的技术培训。专门从事数据库和知识发现的计算机科学家拥有以可扩展的方式查询和分析大量数据的经验，但他们可能不知道与理解社会过程最相关的信息类型。此外，通常有必要改变传统社交网络研究中开发的理论和模型，以纳入在线网络中交互的新特征，甚至开发全新的社会过程模型。要创建和验证这些新模型，需要熟悉用于社交互动建模的社会科学技术，以及对可扩展到数百万甚至数十亿用户规模的复杂网络进行建模和分析的技术知识。该项目汇集了一个跨学科团队，由具有数据库和数据挖掘、网络建模和分析以及社交媒体专业知识的计算机科学家组成，由加州大学圣塔芭芭拉分校的 Divyakant Agrawal 博士领导，开发计算方法来建模和预测社交网络中的许多重要现象：信息该团队正在开发新的算法以及分析和计算工具，可以有效应对大规模的社交网络以及此类网络中产生的数据。这些工具的设计考虑到了人类行为的复杂性，将社会互动的空间、时间和基于关系的方面结合起来。该项目的长期目标是开发有助于更好地理解在线网络中社交互动的工具，开发可靠且可扩展的模型来预测此类社交过程的结果，并创建可以塑造此类结果的应用程序。该项目推进了以下领域的最新技术：海量数据集的查询和分析、复杂网络的建模和分析、社交媒体和社交互动的分析，特别包括多个同时信息传播过程之间的相互作用。这是一项高风险、潜在高回报的研究工作，部分原因是与从社交网络和社交媒体获取和分析数据相关的挑战。为了对社交网络实体和交互进行建模，研究团队需要访问在线社交网络的数据集来构建、验证和验证模型。同样，发现在线社交网络中的信息、知识和用户行为需要访问社交网络数据集。一般来说，由于隐私问题和许多社交网站的专有性质，获取此类数据集是一项重大挑战。这个探索性研究早期概念资助 (EAGER) 项目提供了一套丰富的数据存储库和评估指标，用于开展涉及社交网络和社交媒体的大规模调查。该项目通过从数百万用户的博客帖子和 Twitter 消息中收集大型数据集，以及捕获 Facebook 等社交网络参与者之间互动的适当匿名数据来解决数据挑战。由此产生的数据集和工具的广泛传播将降低具有不同背景和专业知识的研究人员进入社会信息学和计算社会科学的障碍（并降低相关风险）。由此产生的数据集和工具可能会刺激计算机科学中几个分支学科的根本性进步，包括算法设计、网络建模和分析、数据挖掘和知识发现（等等）。该项目为扩大代表性不足的少数族裔和女性在计算机和信息科学领域的参与提供了独特的机会，特别是那些受社会科学现实世界应用（例如理解社会互动）推动的群体。该项目的结果将通过项目网页传播：http://cs.ucsb.edu/~dsl/?q=content/data-driven-framework-analyzing-user-interactions-social-media。

With hundreds of millions of users worldwide, social networks provide incredible opportunities for socialinteractions, entertainment, learning, and political and social change. Hence, there is a growing interest in understanding information diffusion over online social networks. Because many social interactions currently take place in online networks, social scientists have access to unprecedented amounts of information about social interaction. Prior to the advent of such online networks, these investigations required resource-intensive activities such as random trials, surveys, and manual data collection to gather even small data sets. Now, massive amounts of information about social networks and social interactions are recorded. This wealth of data can allow social scientists to study social interactions on a scale and at a level of detail that has never before been possible. However, Social scientists are not traditionally trained in techniques to deal with the massive amounts of data produced by online social networks. Computer scientists that specialize in databases and knowledge discovery have experience with querying and analyzing enormous amounts of data in a scalable fashion, but they may not be aware of the types of information that are most relevant to understanding social processes. Moreover, it is often necessary to alter theories and models developed in research of traditional social networks to incorporate new features of interactions in online networks, or even develop entirely new models of social processes. To create and validate these new models requires familiarity with social science techniques for modeling of social interactions, as well as knowledge of techniques for modeling and analysis of complex networks that can scale to the size of millions or even billions of users.The project brings together an interdisciplinary team consisting of computer scientists with expertise in databases and data mining, network modeling and analysis, and social media led by Dr. Divyakant Agrawal at the University of California-Santa Barbara to develop computational approaches to model and predict a number of important phenomena in social networks: information diffusion, opinion formation, etc. The team is developing new algorithms and analytical and computational tools that can effectively cope with the massive size of social networks and the data produced within such networks. These tools are being designed account for the complex nature of human behavior by incorporating the spatial, temporal, and relationship-based aspects of social interactions. The long-term goals of this project are to develop tools that help better understand social interactions in online networks, to develop reliable and scalable models to predict the outcomes of such social processes, and to create applications that can shape such outcomes. The project advances the current state of the art in: Querying and analysis of massive datasets, Modeling and analysis of complex networks, and Analysis of social media and social interactions, including in particular, the interplay between multiple simultaneous information diffusion processes. This is a high-risk, potentially high payoff research effort due in part to the challenges associated with obtaining and analyzing data from social networks and social media. In order to model social network entities and interactions, the research team needs access to datasets from online social networks to build, verify and validate models. Similarly, discovering information, knowledge, and user behavior in online social networks require access to social network datasets. In general, acquiring such datasets is a significant challenge due to privacy issues and the proprietary nature of many social network sites. This Early Concept Grant for Exploratory Research (EAGER) project provides a rich set of data repository and evaluation metrics for conducting large-scale investigations involving social networks and social media. The project addresses the data challenge by assembling large data sets from weblog postings and Twitter messages from millions of users, as well as appropriately anonymized data that capture the interactions between participants in social networks such as Facebook. The broad dissemination of the resulting datasets and tools will lower the barrier to (and reduce the risk associated with) entry into Social Informatics and Computational Social Sciences for researchers with diverse backgrounds and expertise. The resulting datasets and tools are likely to stimulate fundamental advances in several subdisciplines within Computer Science including algorithm design, network modeling and analysis, and data mining and knowledge discovery (among others). The project provides unique opportunities for broadening the participation of underrepresented minorities and women in Computer and Information Sciences, and especially those motivated by real-world applications in social sciences (e.g., understanding social interactions). The results of the project will be disseminated through the project web pages at http://cs.ucsb.edu/~dsl/?q=content/data-driven-framework-analyzing-user-interactions-social-media.