This document provides an overview of the WeGov project, which aims to help policymakers better engage with citizens using social media. It discusses the need to bridge the gap between what citizens use (social networks) and what policymakers currently use. The project involves building a toolbox of tools to analyze discussions on social networks, extract and inject information, and help policymakers participate in online discussions while respecting privacy and encouraging trust. It outlines the current status of the project and plans for further developing and evaluating the toolbox tools.
1. WeGov Project Overview
eChallenges
Florence, 27 October 2011
Paul Walland & Steve Taylor
University of Southampton IT Innovation Centre
2. Contents
• Project overview
• Background & rationale for project
• Approach of project
– Toolbox
– Best practices for its use
• Current status & next steps
• Demonstration
– Exemplary use case from end users
– Demonstration of toolkit for this case
• Discussion / Q&A
3. WeGov: the project
• Collaborative Research Project supported by the EC
under Framework 7
• 8 partners with expertise in policy making, social
networks, technology R&D, ICT industry
• 30 month project, now at month 22
4. Why WeGov?
• Plenty of talk about government
portals, minister blogs, opinion
collecting web sites…
• But do people really
use them?
6. Bridging the gap
• Allow policy makers to interact directly with
citizens using Social Network Sites
– Use the tools the citizens already use
• Find and understand
people’s opinions
• Become part of the
discussion
• Open dialog
• Respect privacy
• Encourage trust
7. Target Outcomes
• Tools for use by policy makers that:
– Extract information from Social networks
– Inject information into selected Social networks
– Analyse discussions and opinions within social
networks
– Utilise best practice and methodologies
• Technical
• Procedural
• Legal
• Ethical
– Allow policy makers to take advantage of social
networks – the people’s choice!
9. Understanding the
Legal & Ethical Issues
• Policy makers must have confidence that the WeGov toolkit
can be deployed both legally and ethically
• Laws are mandatory, ethics are aspirational
• Citizens must have confidence that ‘big brother’ isn’t
snooping on their private conversations
• The expectation of privacy is different
for different social networks
• Legality isn’t enough on its own
– perception will be critical to the
success of the WeGov toolkit
11. Toolbox
• Toolbox allows the policy maker to:
– Discover opinions from citizens on different
topics
– Visualise and analyse those opinions
– Provide feedback into discussions
– Start discussions
• Toolbox is a framework
– New tools can be added as they become
available
12. Benefits to Policy
Makers
• Automatic scheduled searching of multiple
SNS
• Recommendations of which SNS users to
follow
• Current hot topics
• Find out where people are talking about a
topic
– e.g recommendation of places to start /
contribute to debates
• Summaries of themes and opinions
13. Toolbox
INPUT / OUTPUT
Data
Tools
Policy
Maker
(User)
Execution Control Display
14. Toolbox
• Tools
– “Functions” to get and process SNS data
• Data
– Configuration, input & output for tools
• Tools can be “chained” into Workflows
– One tool’s output can be another’s input
• Execution Control
– Immediate execution (run tool now)
– Schedules
• Automatic, repeated execution
– Monitor runs’ status
15. Tool Categories
Search Find users, posts from
different SNS, monitor groups,
collect data
Analyse
Analysis themes, topics, opinions in
search results, behaviour
Make posts, tweets, etc
Injection in SNS to promote
debate
16. Topic-Opinion
Analysis: Challenges
• Discussions may go across different themes/topics
– We want to recognize them
• Users express different opinions and arguments in their
contributions
– We want to understand them
• Long discussions may contain hundreds of posts
– We want to find out most important ones
• Some users receive higher attention in discussions
– We want to identify them
• Results should be presented in a simple manner,
supporting user navigation through to online contents
17. Topic-Opinion
Analysis: Outcomes
• Topics are latent (sub)themes of interest
– represented by most frequent, characteristic terms
• SNS posts can be classified into topics
– most relevant posts for a topic can be shown to the
user
• Opinions are indicators of emotional states
– (e.g. agreement, disagreement, skepticism..)
• Opinions can be recognized via characteristic
words and phrases („indicators“).
– SNS posts often express opinions
18. Workflow Examples
• Search SNS for
keywords, posts, users
• Analyse search output for
Search Analyse themes, topics and opinions
• Begin a debate by injecting
posts
• Monitor debate
Inject Search Analyse • Analyse debate
• Find a debate by
searching
• Contribute by injecting
Inject posts
Search Search Analyse
• Monitor debate
• Analyse debate
20. Current Status
• Requirements analysed
– Set of exemplar use cases the project should address
• Legal and ethical analysis conducted
– Large impact on project’s decisions
• Now on second prototype
• Toolbox infrastructure mostly complete
– Data store, execution control, dashboard UI
– Plug-in design allows new tools to be integrated as
they become available
21. Current Status
• Exemplary tools in toolbox
– Twitter search
– Socialmention search
– Topic & opinion analysis
• Other tools in pipeline
– Injection
– Discussion activity & behaviour analysis
– Other searches (Facebook, other aggregators)
– News item search (actually any internet document)
• where is it being discussed?
22. Next Steps
• Integrate new tools already developed
• Updates to existing tools
• Complete second prototype system
– Aimed for completion end 2011
• Jan-June 2012: Evaluation in hands of end
users
– Internal to project
– Invited external partners
• New uses and tools for the toolbox!
23. End-Users & Sample Story
eChallenges
Florence, 27 October 2011
Timo Wandhoefer, GESIS - Leibniz Institute
for the Social Sciences
Somya Joshi, Gov2u
24. Policy Makers
GESIS
Criteria: “Decision makers using Web2.0”
• German Parliamentary Party
– Sample topic: German middle class
• Federal Parliament (Germany)
– 3 MPs
– Sample topic: Nature protection
• German Parliament
– 29 MP office members (all parties)
– Sample topics: Nuclear phase-out, web policy
25. Policy Makers
Gov2u
• European Parliament
– Dr. Andreas Schwab
• German – EPP, currently rapporteur for the new consumer
rights directive including EU-rules for on-line shopping
– Mr. Damien Abad
• French – EPP, initiator of MEP 2.0 classes on how fellow
MEPs can use social media platforms like Facebook and
Twitter to reach out to voters
– Mr. Wim van de Camp
• EPP, currently rapporteur for the cleaner and safter
motorcycle regulation
26. Process of
Consultation
• Initial use case & mock-up iteration
– Presentations & discussions
• First prototype “Testing a statement”
– Presentation & discussion at German Parliament
(pre-test for evaluation)
– Modification “Newspaper story”
• Evaluation of the initial
toolbox
– 16 semi-structured
interviews
27. Key Highlights of the
initial Evaluation
• Politicians already
– Post on SNS for public relation purposes
– Monitor particular topics, groups, profiles on Facebook
and Twitter
– Use information services with similar patterns to the
WeGov topic opinion analysis
• Politicians needs
– Finding influential groups & users in order to find
suitable places to disseminate messages
– Engaging with constituencies
– Tools for initiating a dialog with citizens as well as getting
an information advantage in terms of social media
28. Benefits confirmed
by Policy Makers
• Create / modify / schedule workflows
• Disseminate posts / analyse feedback
• Analyse groups / discussions / users / topics
• Setup information feeds
• Use multiple SNS with less SNS knowledge
29. Example Case
“Occupy“
• Across Europe people are demonstrating for
social justice and against the power of bank
• The German chancellor announced her
understanding for the people…
– But does she really understand what drives
people?
• A German Parliamentarian wants to conceive
what people are talking about and what they
think
31. Toolkit Usage for
Case “Occupy”
• This is a “search-analyse” situation
Search Analyse
• We create searches for relevant keywords
– We can schedule automated regular searches
– Gathering data at points in a time period (or
indefinitely)
• We select search results as input for analysis
– To discover themes, opinions and influential posts
– We can select the whole result set or subsets of it