This is the summary of an article by Karissa McKelvey, Alex Rudnick, Michael Conover and Filippo Menczer. It talks about using Twitter to collect huge amounts of data online for a large-scale social study, but make these data and their interpretation easily accessible to social scientists and anyone else interested who may not have the technical know-how otherwise. The paper further describes extensions to the platform responsible for this convenience, called Truthy. Collaboration between persons in the computational and social field can then be further realized. You can get the pdf of the behavioral targeting article here: Visualizing Communication on Social Media: Making Big Data Accessible.
The sociological phenomena that exists in online social networking platforms has been the subject of study over recent years. In particular, it has been found that data from these platforms can be used alongside statistical tools to understand the interactivity and behavior of a huge number of social networking users.
Most of the studies are descriptive in nature, and very few computational techniques are accessible to social science and communication experts. However, it is important to connect quantitative and qualitative measures to further understand social networking systems.
One way to do this is through the use of interactive visualization tools. Ben Schneiderman, a pioneer of information visualization, says that the tool should have the following characteristics: gives a data overview, has filter and zoom, provide detailed data on individual items, show relationships among items, and find a way for target data on specific subsets to be extracted.
This study introduces one information visualization tool for Twitter’s platform. It’s called Truthy.
Truthy Platform
Truthy monitors real-time tweets and clusters them into groups called “memes.” Grouping into memes is done by examining tweets that have common hashtags, mentions, hyperlinks, or phrases or substrings. From this, information propagation dynamics can be obtained in high resolution and in a large scale.
Memes are initially collected into coarse-grained level themes and ranked for ease of navigation. For each theme, users can search for memes and even sort entries using several statistical tools. Each meme is visually represented as a diffusion network, containing features about the number of tweets and users, mean degree and other statistics, and personal user statistics. These and other interface elements allow users to have a detailed look at the activities going on for certain memes. The information presented, however, may be too difficult to understand for experts that are inclined towards qualitatively analyzing social dynamics.
Therefore, this study proposes certain design changes for users to be able to interact with the data, to bridge the gap between quantitative and qualitative studies.
Suggested New Interface Elements
The suggested interface addresses users desire to look at a highly detailed individual-level information, and an overview of the structural positions of these individuals on the meme level. Therefore, the following interfaces are provided: “an interactive layout of the communication network shared among a meme’s most retweeted users and detailed user-level metrics on activity volume, sentiment, inferred ideology, language, communication channel choices, and a realtime feed of each individuals’ recent activity.
Users will expect to find descriptive statistics in the form of total tweets, mentions and retweets, affective sentiment (using OpinionFinder) in relation to the meme, probable language, most recent activity date, date of account creation, and partisan affiliation if United States politics is included.
The new meme diffusion network visualization also shows nodes that are represented by accounts, which, when clicked, reveals further information about that user. The original network is also reduced in this new one to only show the top twenty users, and their neighbors, that are most retweeted.
A Google Chart Tools API based table is also available, which presents the properties used in calculating metrics. These information can be used by researchers to study the demographic and behavioral characteristics of a larger population of individuals in contrast to that for the meme diffusion network.
Conclusion
An interactive visualization platform design was created based on traditional benchmarks and insights from research-active journalism scholars. Users that do not have the technical know how to investigate statistical and computational analysis can use this tool to be informed. This and similar tools will become more and more important as more data is available for online social networking platforms and the research regarding these data rises in demand.
The post Visualizing Communication on Social Media appeared first on Behavioral Targeting Blog.