Human-guided visual analytics for big data.

Abstract

I n this chapter, we provide an overview of the research and practice in visual analytics with a specific focus on decision-support systems that facilitate generating useful information from big, unstructured, and complex data. We first define what is usually referred to as big data and its unique characteristics. We then define visual analytics and human-computer collaborative decisionmaking (HCCD) environments, compare and contrast human-in-the-loop analysis methods with automated algorithms such as machine learning models, and explain how these approaches complement each other for real-world problem solving. To ground our discussions, we provide an overview of four exemplar visual analytics systems with applications in various domains, including humanitarian relief, social media analytics, critical infrastructure vulnerability modeling, resource allocation, and performance evaluation using multidimensional data. MOTIVATION AND OPPORTUNITYAdvanced analytics and computational algorithms enable the transformation of the evolving deluge of digital data into useful and actionable information. However, as data sets continue to increase in size and complexity in the digital

Differentiating geographic movement described in text documents

“…Anselin, Syabri, and Kho (2010) show how providing multiple views of data is important for broad spatial data analysis. Pezanowski, MacEachren, Savelyev, and Robinson (2017), and Karimzadeh, Zhao, Wang, Snyder, and Ebert (2020 utilize multiple views in geovisual analytics interfaces to show geographic data in text documents. Other research into broader visual analytics topics also stresses the importance of multiple views (Andrienko et al, 2020).…”