IHS RISK AND SECURITY Information | Analytics | Expertise AUGUST 2014 SOCIAL MEDIA INTELLIGENCE Practical Strategies for Using Social Media to Enhance Security Dr. Nina Laven IHS Economics & Country Risk Dr. Nathalie Wlodarczyk IHS Country Risk David Hunt IHS Country Risk © 2014 IHS SIC SINGAPORE / AUGUST 2014 • Social media intelligence = SOCMINT Key Message © 2014 IHS • ‘SOCMINT’ cannot be used alone for security. Social media intelligence needs to be paired with other forms of information and intelligence to be effective. 2 SIC SINGAPORE / AUGUST 2014 • What is social media intelligence? Agenda • How to use social media smartly: • Apply Social & Country Expertise • Use Data Science • Create Network Analytics • Emerging SOCMINT Capabilities • Key Takeaways © 2014 IHS 3 What is SOCMINT? SOCMINT involves placing data in context • Social media insights are of limited Data Science Country & Social Expertise Social Media Intelligence (SOCMINT) intelligence value without the context provided by other sources. • Country & Social expertise: expertise on the society and country being assessed, including how people there use social media. Social Network Analytics • Data science: guides the choices we make about how to identify signals in large datasets – which models to use, what samples are valid etc. • Network analytics: patterns in digital social network data. © 2013 IHS 4 What is SOCMINT? Four platforms Social Networking © 2013 IHS Blogs & Microblogs Content Communities Instant Messaging 5 Country & Social Expertise Know your platforms: what matters where © 2013 IHS 6 Country & Social Expertise Know your platforms: how they are used © 2013 IHS 7 Country & Social Expertise: Yemen case study Know your sources: who matters more VALUE • Location • Cultural Proximity • Access • Timing… RELIABILITY • Track record • Endorsement… Source: IHS Country Risk © 2013 IHS 8 Country & Social Expertise Know your sources: how to focus further © 2013 IHS 9 Data science: Twitter data case study Data issues: is your data representative? We want large and valid datasets but we do not always have the resources to purchase and analyze them. How do we find the right balance? © 2013 IHS 10 Data science: Boston Bombing Tweets case study Data issues: is your data trustworthy? Only 20% of tweets had true information about the attacks. Source: IBM Research Labs © 2013 IHS 11 Network Analytics: UK Riots 2013 case study With analytics, social media can aid detection Source: IHS Country Risk © 2013 IHS 12 Network Analytics: UK Riots 2013 case study With analytics, social media can aid detection Source: IHS Country Risk © 2013 IHS 13 Network Analytics: UK Riots 2013 case study With analytics, social media can aid detection Source: IHS Country Risk © 2013 IHS 14 Network Analytics: Limitations Automated sentiment has limitations Source: IHS Country Risk © 2013 IHS 15 Network Analytics: Emerging capabilities How to use the Geolocation of tweets Source: IHS Country Risk © 2013 IHS 16 Network Analytics: Emerging capabilities Social media can map leaderless movements Source: IHS Country Risk © 2013 IHS 17 The strengths and limitations of SOCMINT What SOCMINT can do Limitations of SOCMINT 1 Situational awareness of fast-moving complex developments (e.g. protests, coups) Only a portion of the drivers of global risks are observable through social media 2 Information on events poorly reported in other sources (e.g. due to state control of media) Most data related to networks and communications between individuals is not publically available. 3 Measure the scale and sentiment of online discussion of a topic of interest Algorithms can only go so far in spotting relevant emerging risks 4 Provide timely warning of emergent risks 5 Identify key influencers and groups of users to inform monitoring and outreach © 2013 IHS 18 SIC SINGAPORE / AUGUST 2014 What is the right balance between false positives and false negatives in a system? Questions How do you ensure there aren’t blind spots in your collection approach? The ethics of SOCMINT How can traditional tradecraft co-exist with and enable data analytics insights? How can new analytics tools help the security team build conversations across the enterprise? © 2014 IHS 19