2.1.1 Public Health Surveillance
2.2.1 Limitations of Traditional Data
2.2.2 Opportunities for Social Monitoring
3.2.1 Active vs. Passive Monitoring
3.3.1 General-purpose Social Media
3.3.2 Domain-specific Social Media
3.3.3 Search and Browsing Activity
3.4.3 Social Network Structure
4.1.1 Content Analysis and Filtering
4.1.3 Individual Analysis
4.1.4 Validation
4.2 Qualitative Analysis
4.2.1 Validation
4.3 Study Design
4.3.1 Study Population
4.3.2 Causality
4.3.3 Cross-sectional vs. Longitudinal Analysis
5.1 Disease Surveillance
5.1.1 Influenza
5.1.2 Other Infectious Diseases
5.1.3 Non-infectious Diseases and Chronic Illness
5.1.4 Systems and Resources
5.2 Behavioral Medicine
5.2.1 Diet and Fitness
5.2.2 Substance Use
5.2.3 Disease Prevention and Awareness
5.3 Environmental and Urban Health
5.3.1 Disaster and Emergency Response
5.3.2 Foodborne Illness
5.3.3 Air Quality
5.3.4 Climate Change
5.3.5 Gun Violence
5.4 Healthcare Quality and Safety
5.4.1 Healthcare Quality
5.4.2 Medication Safety
5.5 Mental Health
5.5.1 Depression
5.5.2 Suicide and Self-harm
5.5.3 Mood
5.5.4 Other Mental Health Issues
6.1 Methodological Limitations
6.1.1 Limitations of Self-reported Data
6.1.2 Sampling and Sample Size
6.1.3 Reliability of Third-party Data
6.1.4 Adversarial Concerns
6.1.5 Bias
6.2 Actionability Concerns
6.2.1 Current Use by Practitioners
6.2.2 Reliability of Web Intelligence
6.2.3 Utility of Web Intelligence
6.2.4 Decisions and Interventions
6.3 Ethical Considerations
6.3.1 Public Data
6.3.2 User Interaction
6.3.3 Guidelines for Ethical Research
Preface
In October of 2010, Michael was a new computer science Ph.D. student at Johns Hopkins, Mark was his research adviser, and we were both attending the conference on Empirical Methods in Natural Language Processing