This is an old revision of the document!
What is VizAssist?
VizAssist is a user assistant that helps you turn your data into a visualization. Using your data, some information about your data (called meta-data), and your objectives, VizAssist will look in its list of visualizations to find the most relevant visualizations (w.r.t. your objectives) and for each of them, Vizassist will suggest a mapping from your data to the visualization. You will be able to adjust this mapping in an interactive manner.
What are the main steps of VizAssist
VizAssist proceeds with the following main steps:
Load the data
: Start with a demo file, and in a second step, if you want to use your own data, then please read this Explore your own dataset
Tell VizAssist about your objectives. What kind of data mining tasks do you want to solve? What functionalities would you like to find in the obtained visualizations? If you don't know, then select the “I don't know” option!
Tell VizAssist about your preferences for some data attributes. You can adjust the importance of data attributes. If an attribute is not important at all, then set its importance to 0, and this attribute will not be used in the forthcoming visualization. If an attribute is mandatory, then set its importance to 100. Any other value in between can be used. If you do not want to change the initial importances (usually, all attributes are consideed to have an equal importance), then just click on Next.
Select the visualizations that you favor or that best match your data and your objectives. In this step, VizAssist proposes you a list of visualizations, with thumbnails, and with two scores: the first score is about the matching between your objectives and the visualizations. The second score is about the quality of the mapping between the data attributes (the columns in your table) to the visual attributes of the visualizations. Here VizAssist expects you to select some of the visualizations (or all if you don't know),
Explore the visualizations. In this step, you can use a multiple views interface to explore your data. A liste of preview (with your data) is avalaible on the bottom of the interface. You can click on a preview to obtain a larger view of a visualization. These visualizations might have some interactions that you can use,
Adjust your favorite visualization. The visualization that is enlarged can be further optimized with an Interactive Genetic Algorithm. This search method will propose you a different data to visual signs mappings, i.e., it will change the data attributes that are represented in the visualization by selecting other attributes in your dataset. In this way, you can visually explore the dataset. The IGA has 9 small visualizations. Click on them to obtain a larger view. Select those that you like (tick below the visualization), then click on “Next generation” to obtain new proposals. The IGA takes into account your choice to improve the generated visualizations.