The Challenges and Pitfalls of Aggregating Social Media Data
1. The Challenges and Pitfalls of
Aggregating Social Media Data
Prepared for: NDU Data Summit
27 November
2012
Institute for the Study of Violent Groups
2. Overview
• Social Media Panacea (Don’t believe the
hype!)
• What question are you trying to answer?
• Indications and Warning vs Research/Analysis
• Mexico Case Study
• Understand the media environment in area of
interest
3. Social Media Panacea
• Social Media is a “Silver Bullet”
• Crowd sourced opinion data
• Can provide Indications and Warning
• “Social Media is going to be able to tell me
what is going on the ground in Mexico!”
6. Questions/Insights After the Mexico Study
• What is the data processing time and storage
requirements?
• Was the data pipe size adequate to answer
research question?
• Was the collection profile effective?
• Is their US person data present?
• Who in Mexico is using social media?
• What different demographic groups are using
which social media platforms (FB, twitter, youtube,
etc)?
7. Social Media Next Steps
• MX study left us wanting a country level social
media baseline
• Started looking at traditional media outlets
used SM platforms
• Create an index of traditional media outlets and
their SM handles
• Convene roundtable discussions with media
professionals and academics
8. Hurdles
• There is more disagreement than agreement among experts
about social media, and virtually no expertise about social
media in the developing world
• No academic discipline “owns” social media
• There are no databases or indices available – public or private –
that we could locate
9. Initial Findings
• Gap between what can remotely by collected from Internet
searches and what actually constitutes the media in these
countries.
• Facebook is king
• Twitter is misunderstood in developing countries and generally
disliked; the exception is highly-functioning, usually Western-
educated elites
• Mobile device proliferation is changing social media usage
• Social media, and Facebook in particular, is generally not seen as a
good source of information in these countries, especially when
compared to traditional media
• Most foreign government organizations do not have a solid strategy
for using social media
• Cross-national comparisons are very difficult to meaningfully
interpret; most of the interesting findings are country-specific.
10. Wrap Up
• Be a skeptic and ask hard questions
• Social media is not one size fits all
• Why are you collecting social media data?
– Indications & Warning vs Research/Analysis
11. Questions & Discussion
Carlo Pecori
Program Manager
Institute for the Study of Violent Groups
University of New Haven
cpecori@isvg.org