Cooperative, dynamic Twitter parsing and visualization for dark network analysis

Abstract

Developing a network based on Twitter data for social network analysis (SNA) is a common task in most academic domains. The need for real-time analysis is not as prevalent due to the fact that researchers are interested in the analysis of Twitter information after a major event or for an overall statistical or sociological study of general Twitter users. Dark network analysis is a specific field that focuses on criminal, terroristic, or people of interest networks in which evaluating information quickly and making decisions from this information is crucial. We propose a platform and visualization called Dynamic Twitter Network Analysis (DTNA) that incorporates real-time information from Twitter, its subsequent network topology, geographical placement of geotagged tweets on a Google Map, and storage for long-term analysis. The platform provides a SNA visualization that allows the user to interpret and change the search criteria quickly based on visual aesthetic properties built from key dark network utilities with a user interface that can be dynamic, up-to-date for time critical decisions and geographic specific.

Social Media Exploitation by Covert Networks: A Case Study of ISIS

“…As such, to mitigate this threat, the intelligence community needs to continue to stay abreast of the technological revolution that is social media (Freeman & Schroeder, 2014). Dudas (2013) addresses three areas for future research. The first uses a sentiment lexicon to rate keywords on a positive/negative scale and then maps these keywords against the target keyword over time.…”

“…These networks face constant attempts by governments, the intelligence community, and military groups to disrupt their activities (Milward & Raab, 2006), so their members naturally try to evade detection and intervention (McBride & Hewitt, 2013), sometimes with deceptive or misleading data (Roberts, 2011). The criminal (covert) nature of the organizations behind the networks necessitates quick, nearly immediate real-time analysis and reaction for authorities to successfully counter their activities (Everton, 2012a;Dudas, 2013). Covert networks can be resilient (Milward & Raab, 2006;Everton, 2012b;Senekal, 2014), often overlap with other covert networks (Senekal 2014), and act more like traditional organizations (Raab & Milward, 2003) in their attempts to organize and exhibit structure to their activities.…”

“…They have successfully applied social network analysis (SNA) to covert networks such as corporate price fixing, organized crime, drug trafficking, and terrorist groups (e.g., Krebs, 2002;Sageman, 2004;Xu, Marshall, Kaza, & Chen, 2004;Rodriguez, 2005;Keefe, 2006;Natarajan, 2006;Reid, Chen, & Xu, 2007;Bright et al, 2012;Senekal, 2014). Critical to this success has been the more recent use of social media networks and the analysis of these networks and data using existing social network analysis techniques (Everton, 2012a;Schroeder, Everton, & Shepherd, 2012;Freeman & Schroeder, 2014) and new techniques such as real-time Twitter data analysis and visualization (Cheong & Lee, 2011;Dudas, 2013).…”

Location based popularity analysis of Twitter data

“…The research objective of this thesis is to identify possible ways that can lead us to extract knowledge through the sentiment analysis of location based online social networks. When looking at patterns in large datasets, a popular choice is the utilization of Twitter [23]. Geo-tagged Twitter messages have been used as our primary data source but the data (that our application can analyze) is not limited to geo-tagged tweet only.…”

“…A lower number of tweets were selected for better understanding any trends in popularity. Election prediction has been plotted in the line graph fig 23…”