Segment Drivers

Hardik Chheda Updated by Hardik Chheda

Segment Drivers analyze the change of a particular feature throughout different areas of the data.

Tellius identifies these segments by analyzing the data for a change in behavior pattern in the selected feature vs other features.

Actionable Items:

  • Learn more
  • Top predictive factors
  • Top recommendations
  • Explore Segment

When you select a segment, Tellius displays the selected feature performance in that specific segment for the applied filter conditions.

Segment Information

The Drivers tab displays the segments that Tellius analyses from the data. You can click these segment buttons to view the segment details.

Additionally, the right panel provides the info icon to view in-depth summary about the insights that you have opened.

The segment summary includes the list of features that were considered for your insights and will inform the total data points analyzed, along with the total time it took to process and find your insights.

Key Drivers

This section displays percentage details of top factors in your data affecting the selected feature.

Top Recommendations

In this panel, you can view the top recommended segments of the data according to the feature that you select.

The segments are arranged in an ascending order of the higher percentage they drive the insight.

For each segment, you can explore more specific details, as shown below.

Explore Segment

The Top Recommendations panel provides the Explore Segment button. Click this button to further explore the corresponding segment and get more insights.

You can view the segment details in different formats, like raw data, chart, and smart insights. You can change the chart type from bar chart to bubble chart.

Additionally, the explore segment window provides following options:

  • View Raw Data for Segment Insight
  • Apply the current Segment Insight on other data
  • Add the segment insight to the VizPad page

Note:

  1. Segment Target cannot be calculated on columns due to aggregation as part of it, which means that a calculated column is a measure that is already aggregated. There is no use of segment target variable as its just a single aggregated value.
  2. Row Level calculation will not be considered for Feature Importance.

How did we do?

Drivers

Segment Insights - Model Metrics

Contact