Bing Maps: Penetration Mapping

Added in Q4 2023

Penetration mapping compares particular groups of customers by location. This functionality builds upon the capabilities of the modelling Profile tool which allows you to make comparisons between two groups of people and examine a range of statistics. For example, the profile displayed below compares people who have made a booking to Sweden with the database as a whole, and the Postal Area variable has been added as a dimension of interest. You can see the over- and under-representation in each category very visually in the Penetration column, together with the supporting metrics.

Whilst this provides excellent insight, and allows you to identify categories of significance and interest, it is not possible to display the locations of the postal areas within the Profile tool itself and it is difficult to get a sense of how each area relates to the others.

Instead, within the Map tool, a new Selections option, accessed via the mapping toolbar, lets you specify an Analysis selection, optionally a Base selection, and one of the calculated insight measures which is then displayed for each geographical unit on the map.

Let’s take a look.

  1. Open a new Map.

  2. In the Layer 1 tab, add Postal Area as the geographic variable to use and build.

With no underlying selection, the resulting map displays the number of people on our database living in each UK postal area. The insight is limited. For example, the fact that London contains more people, and mid-Wales less, is linked to the fact that they are more and less populated anyway.

Penetration mapping offers greater insight:

  1. Click Selections on the map toolbar.

The Map Selections dialogue opens with two tabs.

  1. Within the Analysis Selectiontab, drag and drop the selection of Sweden Holidaymakers into the From Selection drop-box.

  2. Switch to the Base Selection tab where you can see that the default is People - i.e. all people on the database. Leave as the default.

At the bottom of the dialogue you can select the calculated insight metric from the Measure To Display drop-down:

  1. Select Analysis Base Index and OK.

  2. Rebuild the map.

You can see that the shading has updated and hovering over a postal area displays the corresponding index value. Here, the index values indicate that the areas around London, Birmingham and Manchester have a higher penetration and mid-Wales lower, suggesting that Sweden holidaymakers are generally located in more urban areas, although there are clearly some outliers, such as the Isle of Man, Glasgow and Hull.

You can add other measures into the map to gain further insight. For example, you might want to compare the average cost of holidays for Sweden holidaymakers against the average cost of holidays for all customers and return the index value on this basis.

  1. Switch to the map’s Layer 1 tab.

  2. Right drag and drop the booking Cost variable onto the Statistics panel and select Mean(Cost).

  3. Make Mean(Cost) the primary statistic and rebuild the display.

You can very quickly and easily discern that the average holiday values for each group demonstrate a very different pattern with, for example, north Wales and the south of Scotland being more over-represented compared to the customer base as a whole.

The available index measures and ratios can be applied to any of the statistics you have listed within the Layer 1 Statistics panel.

 

As well as being able to compare two groups of people within FastStats - i.e. Sweden customers vs. all customers - it is also possible to take records within your FastStats system and compare them against an outside population. For example, you might wish to compare your whole customer base against an overall population count for each postal area. You can achieve this by using a lookup file that has one row for each postal area and the relevant geo-demographic information.

  1. Create a new shaded map with Postal Area as the geographic variable to use.

  2. Click Selections and drag on a selection of All People as the analysis group.

  3. Switch to the Base Selection tab and select From File.

  4. Right click within the Drop your input file here drop-box and select Browse.

  5. Navigate to the saved input file and select.

In this example, a data file supplied with our shape files is used. In the first column it has all the UK postal areas, as well as various age groupings, total population and other geo-demographic information.

  1. Once selected, click Preview to see a sample of the data.

  1. Apply the required settings and select the measure you want to display.

Here, by selecting POP_TOT as the data column, you can compare the count of people in each postal area within your FastStats system against the total population per postal area provided in the data file.

  1. Click OK.

As FastStats takes the total as the Sum(All Rows), if the total is not that then it cannot return sensible results. The following message displays to inform you of this.

  1. Click OK and rebuild the display.

The map now provides completely new insight and it is clear, for example, that customers in Wales and the south west of England are over-represented when comparing all of the Holidays system customers to the underlying populations in each postal area.

The key column in the external data file must exactly match the codes of the corresponding geographic variable in your FastStats system - here, the postal area code.

Whilst the above example is a typical use case and uses a base population from a file - here, population total - to compare against, you can look up any value column so, in other circumstances, it could be an area, some other demographic, a ratio value, or anything that is relevant to the data being mapped.

To provide further flexibility, you can choose to have the analysis from a file and the base from a selection. For example, you might have some sort of multiplier as a ratio from 0 to 1 and want to take your customers and multiply them by this ratio. In this instance, you would use the analysis from file, a base as customers from a selection of all people, and select the Analysis*Base measure.

Though very unusual, you could choose BOTH analysis and base from a file. One such example would be to look into the postal area data file and have the analysis as the population and the base as the area to give a thematic map of the population density in each postal area.