GeoVis - From Terrain to Tweets and Movements

Spatial data exhibits an outstanding value in a broad range of application areas. Waldo Tobler’s first law of geography highlights the phenomenon of spatial autocorrelation: “Everything is related to everything, but near things are more related than distant things.” In other words, the various properties of records of most data sources, such as meteorological data, health data, or even social media data, are frequently dependent variables of the spatial properties. It is thus no wonder that they also play an outstanding role in data visualization. If any feature can be visualized spatially, a visualization researcher will most probably implement a map view if he or she wants to find patterns, trends, and outliers in the data. Every visualization is based on placing graphical marks, such as points, lines, and areas with different properties, such as color, size, orientation, and shape into the so-called spatial substrate – the available space on the screen. Most of the features that can be perceived pre-attentively by a human observer relate to spatial position, spatial patterns, or the form of objects. Therefore, even if the data does not contain any spatial property, say a corpus of text documents, researchers still try to find means to visualize them in a spatial way. With a technique called multidimensional scaling, we can map text data into a two-dimensional pane such that near documents are more related than distant documents.