The Response Analyser - screenshots 4

The propensity modelling toolbox

Screenshots, continued

10. Single variable output

This shows some of the many outputs possible from the univariate analysis. The first table on Book7 illustrates how rarely-occurring values can be combined as "other"; the second table on Book7 shows an age band variable calculated using the variable transformation option.

Book10 shows the rounding option applied at the time of analysis - in this case, selecting "nearest 1000".

If the chart option is selected, each variable is also represented as a pie chart; Book10 has the pie chart for "Var1" on a separate worksheet (you can see the tab.)

11. Cross-tab output

As well as this table of simple counts, as you can see, tables of vertical percentages and of index scores are created. Optionally, there is also a stacked bar chart in a separate sheet.

12. Response table output

This response table output show the response rate according to age bands. Index and z-scores have also been chosen here, as have the highlighting options to show deviations from the norm. By any measure, response is highest for this product among the 65+ age range!

13. Decision tree output

For ease of ilustration, this is a very simple decision tree output, with only three final cells; in a real-life situation, with a large analysis set, the tree can branch several times. Nevertheless, this illustrates how the tree diagram and its associated gains chart are displayed in Excel worksheets, with cell descriptions attached to each cell.

The final node number is attached to the analysis table as an extra variable, making implementation and additional tabulation all the easier - if you want to create comparative profiles of the cells, then it's a matter of moments to run a cross-tab analysis on the data set with the node variable ("Node_1" in this case) as the column variable.