Population Sentiment Analysis
From MilcordWiki
Overview
As part of the Social Network Analysis Reachback Capability (SNARC) project lead by MITRE in support of the Information Dominance Center (IDC), International Security Assistance Force (ISAF) Joint Command, we leveraged our open-source business intelligence platform to perform a multidimensional analysis of survey data from Afghanistan. Our analysis of the sentiment data determines influence patterns within the data to identify significant trends relevant for operations. Integrating tools for population sentiment analysis requires a data strategy to populate capabilities with relevant data.
Need
In a wide-read analysis of intelligence efforts in Afghanistan, MG Flynn writes, “Having focused the overwhelming majority of its collection efforts and analytical brainpower on insurgent groups, the vast intelligence apparatus is unable to answer fundamental questions about the environment in which U.S. and allied forces operate and the people they seek to persuade.” There are a wide variety of surveys collected in operational environments, such as Tactical Conflict Assessment Framework (TCAPF) data collected by military units and others, sentiment surveys, and various polls conducted by non-governmental organizations. Leveraging field collected data to understand the population sentiment is critical in developing more tailored operations that consider the underlying views of the population of interest.
Approach
We have developed a business intelligence suite of applications for our tool to support complex operations management. In business intelligence, the goal is to leverage existing data, determining new insights and relationships to assist in decision making. Our tool uses an open-source online analytical processing (OLAP) approach that provides a multidimensional conceptual analysis of the data. At its core, OLAP is based on cubic structure, consisting of facts that are categorized by dimensions. The cube metadata is created from a star schema, a data warehouse schema consisting of fact tables referencing any number of dimension tables in a relational database. The cube structure of the data allows the analyst to easily navigate through the cube using various OLAP operations such as drilling down/up, slicing, dicing, and pivoting.
Benefits
OLAP provides significant benefits for analyzing data because it is designed to convert data into usable information by allowing a user to break down data into various levels to determine interesting characteristics and relationships. This type of data analysis can have significant impact on operations for an area of operations, in particular for building more effective governance institutions by understanding population sentiments with regards to essential services such as the judicial system and corruption trends.
Applications
- Military: Intelligence, influence and military information support operations
- Civilian: Counter-terrorism campaigns
- Commercial: Brand sentiment tracking
References
- Cassani, L., Caglayan, A., Das, S., Alavedra, J., Morgan, W., and Mooney, L. Population Sentiment Modeling in support of ISAF Joint Command, 2nd International Conference on Cross-Cultural Decision Making in 4th International Conference on Applied Human Factors and Ergonomics 2012 (AHFE 2012), San Francisco, California, USA, July 21-25, 2012. abstract | blog
- Alavedra, J., Stroh, L., Caglayan, A., and Das, S., (2011) Bayesian Analysis of Sentiment Surveys, Fusion 2011, 14th International Conference on Information Fusion, Chicago, IL, July 5-8, 2011.
- Stroh, L., Caglayan, A., Alavedra, J. , and Burke, D. (2011) Population Sentiment Analysis in Support of ISAF Joint Command, HSCB Focus 2011 Conference, Chantilly, VA, February 8-10, 2011.
- Flynn, M. T., Pottinger,, M., and Batchelor,P. D., (2010) Fixing Intel: A Blueprint for Making Intelligence Relevant in Afghanistan, Voices from the Field, Center for a New American Security.
