Correlation visualizer
From Piki
The Correlation visualizer displays the linear correlation between features. The features are spread out along the horizontal axis and the vertical axis shows the correlation with a selected target feature.
| Correlation | |
 | |
| Name | Correlation |
| Can modify data | No |
| Produces filters | No |
| Can show multivariate data | Yes |
| Supports selection | No |
| Supports clipboard operations | Yes |
| Can export data | no |
| Can export images | Yes |
Contents |
Operation
The correlation visualizer will display the linear correlation between one feature (the target feature) and all other features present. The target feature is selected in the feature drop-down selector in the visualizer menu. Negative correlations are shown in red, while positive correlations are colored blue.
To look at the correlation is a quick way to find out what features move together. This can implicate that one of two correlated features is dependent on the other, but it can just as well be that both features are driven by some other ,perhaps unknown, factor. Do not mistake correlation for causal relationship but rather use correlation to get a feeling for the general behavior of a system.
To see the coordinates of the mouse position, right-click on the plot and choose Show World Coordinates to toggle their display.
Zooming and Panning
To zoom to a specific region, use the magnifying glass tool 
. If you just want to stretch or tighten up one axis you can do so by dragging at the axis just outside the plot area (similar to the same action depicted in the Value vs. Sample visualizer). To pan, use the hand tool 
and to get back to the default scales and positions of the axis, right-click and select Original Dimensions.
Clipboard, Save and Print
Their is a dedicated menu for these tasks featuring a button for each action. 
The Save button will save the plot as a any of a number of image formats. Print will let you print out an image of the plot and Copy will copy the correlation matrix to the Windows clipboard in CSV text format. The full correlation matrix will be copied and not just the row displayed.