Social Media for Cities, Counties and Communities

Andrea Kavanaugh, Apostol Natsev, Edward A. Fox, Steven D. Sheetz, Donald J. Shoemaker, Lexing XieVTechWorks 2011

Abstract

Social media platforms like Twitter, Facebook, and YouTube have made a staggering amount of information available to the public and government officials alike. While these resources offer potential for improving services and communication—especially during crises—the sheer volume of data creates substantial "noise" that must be filtered. This report details a 2010 pilot study conducted with government officials in Arlington, Virginia. The study explores how local governments can transform social media data into actionable intelligence, identifying the need for civic-optimized tools and digital libraries to archive and curate content for long-term analysis.

The Challenge: Navigating the Noise

Local governments face a significant paradox: social media provides a wealth of real-time data from citizens, yet the volume of these streams makes them nearly impossible to monitor effectively with limited budgets. Traditional methods of gauging community sentiment, such as mail or phone surveys, are often too slow and expensive for modern crisis response.

The ecosystem of social media communication involving local government and community stakeholders.

A critical gap exists in "civic-optimized" tools. Most social media dashboards are designed for commercial brand management or marketing, not for emergency management or urban planning. Consequently, government agencies often adopt these platforms without a clear understanding of their actual audience, the associated labor costs, or the legal implications, such as public record requirements under the Freedom of Information Act (FOIA).

Methodology

The researchers employed a mixed-methods approach to bridge the gap between social science and data engineering:

  • Qualitative Focus Groups: Officials from Arlington County participated in "electronic brainstorming" sessions. Using group support software, they generated a taxonomy of 23 categories of concerns, split between Organization Factors (policy, legal, cost) and Information Factors (quality, quantity, technology).
Taxonomy of Information and Organization Factors identified through focus groups with government officials.
  • Social Media Analytics: The team utilized video analytics and the Natural Language Toolkit (NLTK) to process data from Twitter, Facebook, and YouTube.
  • Network Analysis: They specifically examined the "extensibility" of communication chains—calculating not just direct followers, but the secondary reach of those followers to estimate potential information "virality" during a crisis.
  • Ontology Engineering: To assist digital libraries in archiving data effectively, they merged four existing disaster databases to create a 185-element "CTR ontology" for the Crisis, Tragedy and Recovery Network.

Key Insights and Results

The Power of Indirect Reach

The most striking insight from the study is the scale of the "follower-of-follower" network. While 34 civic organizations had approximately 31,000 direct followers, those followers reached over 67 million people. This suggests that a single tweet from a local agency has a massive theoretical reach if retweeted by its most influential followers.

Engagement and Categorization

  • Facebook: Traffic was the most discussed topic (approx. 30% of posts), followed by Public Service Announcements (PSAs). While traffic posts were frequent, "Miscellaneous Events" and "News" generated the highest volume of public comments.
  • YouTube: Researchers successfully clustered 1,800 videos into six civic themes: Law Enforcement, Transportation, Social Issues, Economic Development, Political, and Communication.
  • Ontology Development: The team produced a draft ontology of 185 elements, providing a hierarchical structure for classifying various types of disasters and emergencies.

Social Media as a Strategic Asset

In an era of shrinking government budgets, social media acts as a low-cost alternative to traditional community outreach. By monitoring tag clouds and engagement patterns, officials can gain real-time insights into the "mood" and concerns of the community that were previously difficult to capture.

Limitations and Future Directions

The study identified several hurdles that remain for effective government use of social media:

  • Contextual Challenges: The brevity of platforms like Twitter (then limited to 140 characters) makes it difficult to accurately identify context and meaning, leading to potential misclassification.
  • Real-Time Alerting: While post-hoc analysis is possible, building systems that alert officials to meaningful patterns in real-time remains a significant technical challenge.
  • Demographic Bias: Social media may miss older populations or those who cannot afford the technology, potentially leaving certain neighborhoods "falling through the cracks."

The report concludes by outlining the need for more sophisticated multimedia text mining and broader studies to include household surveys, ensuring that digital tools serve all members of the community effectively.

Summary updated

Abstract

Social media (i.e., Twitter, Facebook, Flickr, YouTube) and other tools and services with user generated content have made a staggering amount of information (and misinformation) available. Some government officials seek to leverage these resources to improve services and communication with citizens, especially during crises and emergencies. Yet, the sheer volume of social data streams generates substantial noise that must be filtered. Potential exists to rapidly identify issues of concern for emergency management by detecting meaningful patterns or trends in the stream of messages and information flow. Similarly, monitoring these patterns and themes over time could provide officials with insights into the perceptions and mood of the community that cannot be collected through traditional methods (e.g., phone or mail surveys) due to their substantive costs, especially in light of reduced and shrinking well as community organizations, businesses and the public. We were especially interested in budgets of governments at all levels. We conducted a pilot study in 2010 with government officials in Arlington, Virginia (and to a lesser extent representatives of groups from Alexandria and Fairfax, Virginia) with a view to contributing to a general understanding of the use of social media by government officials as gaining greater insight into social media use in crisis situations (whether severe or fairly routine crises, such as traffic or weather disruptions). Our findings from the pilot study fall into three main areas: 1) local government agencies use social media without knowing its costs and benefits, or who their actual audience is, who in their organization should monitor communications, how and when they should be responding, and what effect their social media communications have on the public; 2) new tools are needed to help government and citizens make sense of the overwhelming amount of data that is being generated, to model the flow of information, and to identify patterns over time; and 3) digital libraries are needed to archive and curate generated content, especially for crisis and social convergence situations, but also for analyses that cover longer time frames. Our work also builds on a larger set of studies carried out by a Virginia Tech team funded by NSF (IIS-0916733) with the overarching goal of building a Crisis, Tragedy and Recovery Network (CTRnet) (http://www.ctrnet.net).
BibTeX
@inproceedings{kavanaugh2011social,
  author = {Andrea Kavanaugh and Apostol Natsev and Edward A. Fox and Steven D. Sheetz and Donald J. Shoemaker and Lexing Xie},
  title = {{Social Media for Cities, Counties and Communities}},
  booktitle = {VTechWorks},
  year = {2011},
}
Figure