Exploring and Visualizing the History of InfoVis
Daniel A. Keim, Hernández Barro, Christian Panse, J. Schneidewind, Mike Sips · 2004
The exploration and visualization of large information spaces is a challenging task. The provided contest data set for example contains more than 1000 authors of about 600 papers. The basic idea for an effective data exploration is to include the human in the data exploration process and com-bine the flexibility, creativity, and general knowledge of the human with the enormous storage capacity and the compu-tational power of today’s computers. The key concept of our visualization approach is to visualize the whole dataset to provide a first overview and to provide rapid, incremental, and reversible analysis actions. Our techniques follow the well-known Information Seeking Mantra: overviews first, zoom and filter, and details on demand. All visualizations, actions, and details are tightly coupled using the well-known linking and brushing concepts. 2 Data Cleaning When processing and visualizing large data sets, data clean-ing as part of data pre-processing is a very important step, since it directly influences the quality of the visualization. Since there where some inconsistencies in the contest data set, like ambiguous authors or different formats and spellings for the conference names, some data cleaning was necessary. Therefore we wrote some shell scripts, based on regular ex-pressions, to correct these inconsistencies. Additionally we corrected the spelling of some author names manually. An-other problem was, that for several attributes no values were