The following sections are included on this page:

Instructions on setting up variables and weights for NPS available here…

Notes on creating the analyses supplied

These notes show how to calculate the NPS

  • by using a cell reference calculation in the doughnut chart
  • by using a weight to calculate the NPS against another variable to give the NPS over time

Creating the NPS doughnut chart in your own survey

This example assumes that you have created a derived variable (V1aNP) that splits the responses into detractors, neutrals and promoters. Doing this is described in Creating the derived variable.

  1. Click charts to open the Analysis Definition dialog.
  2. Change the Chart style to NPS Doughnut Percent.
  3. Enter V1aNP (or the name of your derived variable if it is different) in the Analysis field.
  4. Leave the Calculate field as Counts & Percents.
  5. Clear the Counts box under Show options.
  6. Select the Analysis Percents box.

    Analysis Definition_chart_nps

  7. Change to the Notes/Titles tab.
  8. Enter NPS doughnut in the Title pane.
  9. Click in the Notes pane.
  10. Click [Insert] and select Cell value field from the menu.
  11. Enter (R3C1-R1C1) as the cell value expression. This calculates the percentage of promoters – the percentage of detractors using the values in the current analysis.
  12. Click [OK].

    Analysis Definition_notes_nps

  13. Click [OK]. The chart window opens, displaying the NPS doughnut. If you have no data in your survey, there will be nothing to display in the chart. To see a chart with data, open the NPS sample survey, click Results window button to open the Analyses window and double click AN1 to open the sample NPS doughnut chart.

Creating the bar chart of the NPS over time in your own survey

  1. The chart assumes that you have created the derived variable V1aNP and a weight as described in Creating the weight. It also assumes that you have created a variable used to categorise responses by date.
    An example derived variable Vmonth is supplied in the example NPS survey. You can see how the variable has been defined by clicking codes to open the Variables window, scrolling to Vmonth and double-clicking it to open it.
  2. Click charts to open the Analysis Definition dialog.
  3. Change the Chart style to Horizontal Bar Percent Transposed (or use the nps_blue_bar style provided in the supplied survey).
  4. Set the Analysis field to your derived date variable.
  5. Set the Calculate field to Means & Significances.
  6. Set the mean to be calculated to WT1 (V1aNP). This is the derived promoter variable weighted to give negative and positive values.
  7. Select the Transpose box.


  8. Click [OK] to display your chart. If you have no data in your survey, there will be nothing to display in the chart. To see a chart with data, open the NPS sample survey, click Analyses to open the Analyses window and double click ANyear.2 to open the sample chart of NPS over time.

Notes on using the chart styles supplied

These notes show how to export and import chart styles. The sample survey uses special chart styles to create its analyses. These have been supplied as separate .csf files.You can use these styles in other surveys.

Exporting a chart style from a survey

  1. Click Analyses to open the Analyses window and double-click the chart that you are interested in to open it.
  2. Right-click the chart to open the context menu and select Save Style.


  3. A Save dialog appears. Give the chart style a name and choose a folder to save it in.

Importing a chart style

  1. Place the chart style file (.csf) in a known place.
  2. Open the chart you are interested in.
  3. Right-click the chart and select Load Style from the menu.
  4. Browse for the .csf file that you want.

You can also find a saved chart style by clicking the [Browse] button on the chart definition dialog.
RD: definition for combination chart

Creating your own image bar charts

If you wish to create your own image bar charts, follow the points below to change the stacked bar chart style supplied with Snap to an image bar.

Notes on creating the analyses supplied

The sample survey shows a stacked bar chart made up of a line of people. This can be done quite easily but has to be carefully sized for it to look clean. You will need an image that has been duplicated with a different colour for each element of the chart. The example in the survey used three versions of the image of a man, one in red, one in yellow and one in green.

  1. Click charts to open the Analysis Definition dialog.
  2. Change the Chart style to Horizontal Stacked Bar Percent Transposed.
  3. Enter your variable in the Analysis field (V1aNP in the example)
  4. Select your analysis option (Analysis percents in the example).
  5. Click [OK] to create a stacked bar chart.
  6. Double click on one of the bar sections to open the chart designer for that series.
  7. Select Datapoint defaults in the left-hand pane. (If it is not visible, click Expand branch square icon by Series, then Expand branch square icon on the section title to display it.)
  8. Select the Picture tab on the right-hand side.
  9. Click [Browse] and browse for the appropriate picture.
  10. Select Embed picture to store your picture within the survey.
  11. Click [Apply] to display the change on the chart.

    Chart Designer_picture_man

  12. Select Datapoint Labels on the left-hand side.
  13. Select the Appearance tab on the right.
  14. Set the Text Location to None to remove the percent figures displayed on the bar.
  15. Click [Apply] to display the change on the chart.

    Chart Designer_no_label

  16. Repeat for the other points in the series (Detractors, Neutrals and Promoters in the example.)
  17. Click [OK] to close the chart designer.
    Adjust the window size until the scale of the images is appropriate for the numbers. To fit ten people in, you can use a window size of 841 * 460. The example in the report uses a window size of 900 * 446 to make the one figure represent 10% in the space available.
  18. Click save to save the chart.

If there is a topic you would like a worksheet on, email to