Tree grid

Different views¶

The Tree Grid displays the numeric details behind each node. This view can also be used to see the implications of the Marketing Costs.

The following four alternative views are available, each having its particular advantage.

Hierarchical

Enables you to drill down, expanding and collapsing nodes to see the structure of the tree.
Flat - All nodes

Displays all nodes from all levels in a single table.

Useful for sorting the nodes by a particular column and for making selections at any level.
Non-Cumulative Gains Table (Flat, Leaves only)

Displays just leaf nodes from whatever level of the tree in a single table.

Useful for sorting the nodes by a particular column and for making selections o fleaves.

Activate by just clicking Leaves only.
Cumulative Gains Table (Flat, Leave only, Cumulative)

Displays just leaf nodes from whatever level of the tree in a single table. By default is sorted by descending gain. This is also the order in which the counts and percentages are cumulated.

Useful in conjunction with the Gains Chart to see the total number of records selected in the top N nodes.

Activate by just clicking the Cumulative button.

These views are selected using the buttons at the top of the grid.

Notice that some background logic enforces the above 4 views. For example, if you select Cumulative, the Flat and Leaves Only buttons will automatically be selected too, since it is meaningless to view cumulative statistics with non-leaves included.

Sorting¶

To sort by a particular column, right click on the column header and select “Sort Ascending” or “Sort Descending”. You can only sort by one column at a time.

Example tree grid statistics¶

The above columns from the Tree Grid can be summarised as follows:

Node Id: Unique number for this node, based on the order in which it was created.
Level: The node is created from the root node, using just 1 split.
Rule: The full rule defining people is those with an income of “Unknown or £20k - £80k”.
Base Count: There are 304,326 people in this node (i.e. who satisfy the rule):

% of [all] Base: These people make up 26.31% of the total base selection. - Analysis Count: There are 17,463 people in this node and in the analysis selection:

% of [all] Analysis: These people make up 69.37% of the total analysis selection

Analysis %: These people make up 5.74% of this node. - Gain: This shows how well we have homed in on our analysis selection, or in particular how the Analysis % in a particular node, compares to that in the root node:

In the focus node above, the gain is 2.64 (=5.74 / 2.18). This node contains a disproportionate amount of people from the analysis selection.

A gain of 1.0 would be in line with the root node. - Index: This is another measure of how well we have homed in on the analysis customers, by looking at the ratio of the % of Analysis to the % of Base:

An index of 100 is in line with the root node

In the focus node above, the index is 273 (=100*69.37 / 25.36)

For more details on the statistics, see the section Statistical Terminology.

Available columns¶

Within each of these displays, the following columns are available by right clicking on the headings and using the Column Chooser:

Node Id - The unique number of this node, based on the order in which it was created.
Level -The level of this node (root = level 1).
Gain - The ratio of this node's Analysis % to that of the root node.
Rule - The values of the dimensions that define this node.
Stopping Reason - Why this node has been stopped from growing (if at all).

In addition the following statistical information is available (for a definition of these terms see Statistical Terminology.

Analysis Count
Non-Analysis Count
Base Count
Analysis %
Non-Analysis %
% of [all] Analysis
% of [all] Non-Analysis
% of [all] Base
Index
Z-Score

The following profitability information is also available (for a definition of these terms see How do I set the Marketing Costs?

Cost
Revenue
Profit
Revenue ROI
Profit ROI