Filter combinations for comparing "Last two Instances"

Datasets offer an easy and unuitive way to refine data when comparing history instances. This way, it is easy to know which rows have been added / deleted or changed. For introductary information on Snapshot Datasets and comparing data instances, refer to: Snapshot Datasets: Comparing "Last Two Instances" / Track Changes.

Additionally, to narrow the data down to specific requirements / parameters, users can build sophisticated filters. This articles gives basic understanding on how to use filters when comparing instances.

This article answers the following questions:

  1. How to compare specific values for current and prior period (for example, compare sales made via a specific channel yesterday and day before yesterday).
  2. How to add simple filters
  3. How to add compound filter criteria

1. Comparing current and prior values for a specific field

  1. Filters allow defining additional criteria to compare current and prior values by various conditions. NOTE: Current and prior filters are shown only in the Track Changes mode.
  2. Results: Sales and Prior Sales columns are shown in the Results set and only those rows where the condition is met are shown.

2. How to add simple filters

Use Case: Include rows if the sales in any country have increased by any amount over the prior day.

  1. For each fields: Select the field with constant (or unchanged) values. This field will serve as a basis for comparing other changed values. In our example such fields are Channel and Country (i.e. you are looking for changes when Channel and Country are unchanged)
  2. Click [+ Rule].
  3. Construct a Rule by selecting parameters for comparing current and previous values from the drop-down lists. You may add as many Rules as required tp seed put the unnecessary data.
  4. Apply changes to update the Results set
  5. Results: Sales and Prior Sales columns are shown in the Results set and only those rows where the condition is met are shown.

You may create a separate View out of this data slice by clicking Save as View at the top of the page.

3. How to add compound filter criteria

Use Case: Include rows if units sold via 'website visit' channel have increased over the prior day OR if sales made via 'website visit' channel are 10% higher than over the prior day.

You want the key fields across which changes will be tracked included in the Select Field list.  In this example, you should select ONLY Channel, Country, units, and sales for inclusion in the display.

  1. For each field: Select the field with constant (or unchanged) values. The Id fields should be Channel and Country (i.e. you are looking for changes when Channel and Country are unchanged)
  2. Define filters: Then you can apply filters. In our use case, this consists of 2 groups of conditions: click [+ Group] and define criteria for the first and second group:
    • [Group 1]: units increased in 'website visit' channel
    • [Group 2]: sales are 10% higher in 'website visit' channel
  3. Choose OR filter to define relations between the Groups
  4. Apply changes to update the Results
  5. Results: Units / Prior units and Sales / Prior Sales columns are displayed in the Results set and only those rows where the condition is met are shown.

You may create a separate View out of this data slice by clicking Save as View at the top of the page.

0 Comments

Add your comment

E-Mail me when someone replies to this comment