Automated Visualization

This Component automatically visualizes data in a dashboard based on the most interesting statistical properties of the input columns. A detailed explanation is available for each chart inside the dashboard. No configuration is necessary in order to automatically create the dashboard via the Component Composite View.

The Component runs a various statistical tests on the input data. Candidate columns are ranked for each visualization. The columns with highest rank are, when possible, visualized in a chart dedicated to precise statistical test.

Statistical tests used:

- Skewness, Kurtosis and Spikeyness tests of numerical column distribution
- Correlation test between numerical columns
- Correlation test between categorical columns
- Correlation test between numerical and encoded categorical columns
- Correlation test between categorical and encoded numerical columns
- ANOVA test (between categorical and numerical)

At the bottom of the dashboard you can also select the columns to be removed from the data set.

Input ports

1. Type: Table
   Input Data to Visualize
   Unexplored data to be visualized.

Output ports

1. Type: Table
   Output Data
   Input data with only the columns that were not removed in the interactive view.