VisLink: Revealing Relationships Amongst Visualizations

Barbara Fritsch
Wolfgang Knecht


Generally datasets are stored in textfiles as tables where the semikolon (";") is used as the column separator. A line break indicates a new row. The first row contains the titles of the columns and the second row stores information about the type of data in the respective column. The additional rows contain the data.

The following types of data are valid:

  • string
  • float
  • percent (indicates that it's a float, but the scale for some visualizations has to go from 0 to 100)
A special case is a column with the title "State". When loading the dataset and a column with the title "State" occurs the loader assumes that the following columns contain additional information about the state. The additional informations are the short form of the state and it's coordinates on the map.

Example dataset

Our example dataset is about the election in the USA. The dataset is structured as follows:

  • State (string): names of the states
  • short (string): short names of the states
  • x (float): x positions of the states on the map
  • y (float): y positions of the states on the map
  • EV (float): number of the electors
  • Obama (percent): election results for Obama in percent
  • McCain (percent): election results for McCain in percent
  • EV Obama (float): electors voting for Obama
  • EV McCain (float): electors voting for McCain
The example dataset can be download here: data-election.txt