Geo-social visual analytics has been studied by Luo and MacEachren , who have found that in social network analysis, spatial data is treated as background information and in geographical analysis, network analysis is oversimplified. Hence, they have argued for integrating knowledge from both analysis for datasets which have both geographic as well as social network data, e.g. location based social networks. They have called this integrated analytical approach as geo-social visual analytics. In their work, they have proposed the theoretical framework where the contexts can be brought together, and they have reviewed the state-of-the-art in the three core tasks for geo-social visual analytics, namely, exploration, decision making, and predictive analysis, to find gaps. They have used this gap analysis to identify potential challenges and research directions. Here, we discuss the authors’ views on the conceptual framework and the core tasks of data exploration and decision making.
- The First Law of Geography says, “Everything is related to everything else, but near things are more related than distant things,” and the social network analysis dogma says, “Actors with similar relations may have similar attributes/behavior.” The authors have discussed how the First Law of Geography and the social network analysis dogma must be both combined to define geo-social relationships. At a conceptual level, these relationships are considered to be the intersection set of three different kinds of embeddedness. The different kinds of embeddedness are namely societal, network, and territorial. Thus, the conceptual framework for geo-social relationships states that “Nearness can be considered a matter of geographical and social network distance, relationship, and interaction.’
- The core task of data exploration must consider the relative strength between the two aspects of the data. One of the aspects qualify geo-social relationships as being “among geographical areas” and the other, as being “among individuals at discrete locations.” While the visualizations generated are similar for both sets, for the latter,it requires additional computational processes for exploring the “spatial-social human interactions focus on developing quantitative representations of human movements.”
- For the core task of decision making, spatial data analyses exploit spatial locality to study trends, and relate them to explanatory covariates such as demographic data. Network analysis grows in a bottom-up fashion, whereas geographical analysis tends to be top-down. Integrating both can be done using linked methods where individually the visualization techniques cater to specific needs of both the analyses.
Some of the future research directions are towards:
(a) developing theory, methods, and tools, integrating the two separate aspects of the data;
(b) understanding the dynamics of geo-social relationships and processes;
(c) integrating ideas in cognitive sciences supporting geo-social visual analytics; and
(d) developing new geo-visual analytical methods for the three core tasks.
Virtual Reality for Geovisual Analytics has been studied by Moran , who have proposed that one could separate the preprocessing step of data modeling and analytics from the visualization processes. The visualization process can then be implemented in a virtual reality platform so that the user gets better situational awareness and can use natural user interactions for further analysis.
While the visual representations are approximately the same in immersive as well as non-immersive visualizations, we find that the user interactions for analytical tasks are improved in immersive geovisual analytics. The five tasks in the visualization workflow include navigation or exploration; identification and selection; querying and filtering; clustering; and details-on-demand. Navigation is achieved through fly-through cameras with change of perspectives, where the user can virtually navigate through the life size virtual model of the space. Depending on the approach of the user to objects in the scene, zooming can also be used.
Identification exploits the differences in rendering different geometric primitives for the visual representation of the data, e.g. shape and color can be used for “picking” features of interest. Filtering and querying can be provided using a virtual keyboard, and menu options in the Graphical User Interface (GUI). Clustering and pattern matching can be done by virtually overlaying or stacking different layers pertaining to the data, e.g. height of the data with population. Details-on-demand can be implemented using the levels of organizing data, and drilling down deeper as per requirement. In an immersive virtual reality, the user can a directed 3D arrow to enable direct interaction with the scene and enabling movement as required.
 The age of the area of geovisualizations is not certain. There are references to visualization in paper as early as Philbrick (1953). This observation has been made by MacEachren and Kraak (1997).