Positional wildcards

The Pattern Match feature of on the fly aggregations supports the following positional wildcards:

Wildcards allow you, for example, to select people who bought product A, then another product, and then product A again

With the use of the positional wildcard ?N (where N is a number) at any place in the pattern, it is now possible to refer to other items within the pattern match sequence and say that they are explicitly the same as, or different to, others within the sequence.

When FastStats first encounters a pattern - ?1 – it stores that value, e.g. Destination of United States. Subsequent references later in the pattern will always return the same value. Any other ?N must, therefore, match a different value - i.e. ?2 in this example could not be United States.

Example pattern match sequences:¶

Australia - ?1 - Australia - ?1 - Australia

This will select and return people who have been to Australia, then somewhere else, then back to Australia, then back to the same somewhere else, and then Australia.

Or:

?1 - ?2 - ?3 - ?2 - ?1

This will select and return people who have been to Destination 1, then Destination 2, then Destination 3, then back to Destination 2 and back to Destination 1.

When manually defining a pattern, the pattern elements you can use to create sequences on a value (numeric or currency) variable are as follows.

Using booking Cost as the value variable in these examples:

• Fixed numeric values - F200 to find transactions which are exactly £200 in value.
• Fixed numeric ranges - F>100, F<1000, F500-800 to find bookings which cost more than £100, less than £1000 and between £500-800 respectively.
• Relative numeric ranges - R10-30 to compare transactions and look for a transaction that is between £10-30 more than the previous transaction.
• Percentage difference numeric ranges - P10, P10-20 to calculate that the next value in the list is 10%, or 10-20% bigger than the previous transaction.

Example 1¶

1. Open a new Expression window and click on the Add Aggregation icon.

   To access the ‘on the fly’ options:

2. Click onto the Frequency(Bookings) tab.

3. Select Pattern Match from the Type drop-down menu options.
4. Set the Grouping Table to People and the Transactional Table to Bookings.
5. Order records by Booking Date.
6. Add Destination as the Pattern Match Variable.
7. Click Set Pattern to open the Define pattern for Destination window.

8. Click on the drop-down menu in the Value 1 cell to access the available options.

   

9. Select ?1.

   Note

   Wildcard numbers must be introduced into the pattern in order. Once selected, when you subsequently click onto the Value 2 cell, the options to select ?1 and ?2 will be available, etc. It is not valid to enter ?1 - ?3 - ?5.

10. Complete the settings and enter the Pattern Name as in the screenshot below.

The pattern matched for each person will be dependent upon the value recorded against ?1. The screenshot below demonstrates this:

• Person URN 826137: Italy-Portugal-Italy
• Person URN 796445: United States-Australia-United States
• Person URN 792550: United States-Portugal-United States

These wildcards can be used with:¶

Fixed values

e.g. ?1 Australia ?1

Returns: Destination 1, then Australia, then back to Destination 1

= wildcard

e.g. ?1 = ?2 =

Returns: Destination 1, then Destination 1 again, then Destination 2, then Destination 2 again

? wildcard

e.g. ?1 ? ?2

Returns: Destination 1, then anywhere (that could be Destination 1), then Destination 2

* wildcard

e.g. ?1 * ?2 * ?1

Returns: Destination 1, then any number of Destinations (that are all Destination 1), then Destination 2, then any number of Destinations (that could be Destination 2), then Destination 1

Example 2¶

You can use [ and ] as wildcards to represent the beginning and end of a person's transactions. This allows you to specify whether the desired pattern should be anchored to the start or end of the transactions you are analysing - or in some cases, both.

Let's imagine that you are interested in identifying people whose first three holiday bookings are to Australia then the United States and then Greece.

1. Repeat steps 1-8 in example 1 above.

2. Name and define the pattern for Destination as shown below.

   

3. Name your expression.

4. Drag the expression onto a new selection and enter the pattern name.

   

5. Drag and drop a data grid onto the selection. Add Destination and Booking Date (ordered ascending) and build the display.

   

You can clearly see that each person's transactions start with the defined pattern of Australia, then the United States, and then Greece - even if they then go on to have further bookings to any other destinations.

You can also:

• Use the ] wildcard to anchor the pattern to the end of a list of transactions.

  

  

• Use both [ and ] to search for people whose whole sequence of transactions matches your defined pattern.

  

  

Summary¶

The Pattern Match feature of on the fly aggregations supports the following positional wildcards: