Whoops…Nothing found

Try other keywords in your search


 1 Minute



What Is It?

Insights are automatically generated observations in plain English that are available for certain plot types. Statistically significant regions of data are presented to the user as Insight Cards, which can be easily saved as an Annotation or copied into a presentation.

Why Is This Important?

Insights are useful for minimizing time to value on datasets where exploration and discovery are the main tasks. Additionally, Insights can call your attention to regions of obscured data points while working with big datasets. For example, Insights can highlight an area of the plot where specific values for key metrics lead to higher shipping delays. Insights can also identify important combinations of characteristics that lead to higher spending, providing a quick and powerful way to perform customer or product segmentation.


Insights in Explore are available for Scatter Plots, Histograms, Line Plots, and Network Graphs. 

Open the Insights panel by clicking the icon in the toolbar or selecting Insights from the Data Analytics menu.

Scatter Plots

Insights are generated for Scatter Plots to highlight regions of the graph that show statistically significant distributions of points based on their Color. Bar charts are used to highlight the breakdown of points within each region of interest.


Insights are generated for Histograms if the software detects a significant distribution. In these situations, Insights identify bins that have interesting behaviors based on the breakdown of the colored points that fall within that bin.

Line Plots (Time Series)

Insights for line plots are generated when a mapping has been made in Line Plot mode with distinct lines. These Insights are driven based on what has been mapped to the X-axis (for example, a column containing dates or times would work well) and show all the groups of lines that are correlated. The dropdown list can be selected to offer a breakdown per line and highlight which lines are either correlated or anti-correlated. The algorithm uses normalized cross-correlation (NCC) with a threshold of greater than 0.9 to determine whether two lines are correlated.

Network Graphs

Explore offers specific insights to help understand network data. Please refer to our documentation pages on Network Insights.

Was this article helpful?