0 ×

Geographic Shredder

Market Simulation nodes by Scientific Strategy for KNIME - Community Edition version 4.3.1.v202102181051 by Decision Ready, LLC

The Geographic Shredder node compares the Geographic Location of each Customer against the Location at which each Product is available. It will then shred those Products that are not available from the Customer's Consideration Set by setting their WTP to zero.

The Location of each Customer can be pre-defined upstream by setting the 'Location' column of the Input WTP Matrix. If this 'Location' column is missing, the node will first generate this column by placing each Virtual Customer in a geographic 'Location' according to its 'Population' as defined in the 'Input Customer Locations' table. Customers can be located in different cities across the country, or in different districts within a city. Customers can also be Located at the different online Location they prefer to shop. Customers can be systematically located in blocks, or can be randomly scattered across all locations according to the relative populations at each.

A Simulate Market node assumes, by default, that all Customers have access to all Product offerings. But this usually does not reflect the reality of the Market. A real-world Customer may never travel to an inconvenient Location to buy a Product as the 'Lost Value' may be prohibitively high, or the Customer may not be aware of that Product's existence.

Note that the Geographic Shredder node disregards the Customer's first Top Choice. While the Demand Shredder node will always retain the Customer’s Top Choice, the Geographic Shredder considers only the Location of the Customers and Products. By shredding the Customer’s Top Choice, the results of an upstream Tuning Algorithm can be significantly impacted by a Geographic Shredder node. A downstream Simulate Market node will yield different results after the WTP Matrix has been shredded than before.

Hence, unlike the Demand Shredder node, the Geographic Shredder node is best located upstream of a Market Simulation Tuning node before the final WTP Matrix values have been calculated.

Alternatively, the Demand Shredder node can be used to shred Products by Location. This guarantees the Customer will be able to purchase their Top Choice, and allows the node to be located downstream to preserve the upstream Tuning of an Input WTP Matrix. This also allows the Customer to shop at several Locations. But this does not guarantee the Population of each Location.

The Geographic Feature node is another alternative. The Geographic Feature node calculates the 'Lost Value' of each Product due to shipping cost and delay from the perspective of each Virtual Customer. This gives each Customer a wider Consideration Set, but penalizes the value of Products sold at inconvenient Locations.

More Help: Examples and sample workflows can be found at the Scientific Strategy website: www.scientificstrategy.com.


Standard Options

Locate Customers
The method by which the Virtual Customers will be assigned a Geographic Location. These options will be ignored if the Input WTP Matrix already contains a "Location" column. The options include:
Randomly (default) scatter the Customers across each of the 'Locations' listed in the 'Input Customer Locations' table according to the 'Population' ratio. As Locations are drawn randomly there is no guarantee that the output 'Population' ratio will precisely match the input ratio. If precise ratios are important, use the 'In Large Blocks' option instead.
In Small Blocks placing the actual 'Population' number from the 'Input Customer Locations' into the Customer Distribution and looping over the input table until the Customer Distribution is full. This option depends upon the 'Population' field in the 'Input Customer Locations' being expressed in integers that sum to a fraction which evenly divides the total number of Customer rows in the 'Input WTP Matrix'.
In Large Blocks builds an ordered Customer Distribution table according to the 'Population' ratio in the 'Input Customer Locations'.
Save Randomizing Seed
A Randomizing Seed can be saved to ensure that the Customer Distributions are always shredded in the same way. If a Demand Shredded node is copied then the user should ensure the Saved Randomizing Seed is changed or not saved - otherwise demand may be inconsistently shredded. The 'New' button will generate a new Randomized Seed. Disable the CheckBox to generate a new Randomizing Seed each time the node is run.

Input Ports

Input Product Array: (optional) The set of Products that define the Market. Each row corresponds to a Product that competes for Customers in the Market. The 'Product Array' must have the following columns:
  1. Product (string): The unique name of the Product corresponding to a column of the same name in the 'Input WTP Matrix'. There can also be an additional row with a Product named 'No Sale' - this row is used to handle those Customers who are in the Market but have not yet purchased a Product.
  2. Location (string): (optional) The name of the Product's default Location. These Locations will automatically be added to the 'Input Product Locations' table. If the Product is available everywhere to all Customers then this value can be set to 'Global'.
Input Product Locations: (optional) The 'Location' of each point of sale for each 'Product' in the Market. If this table is missing then only the 'Location' field within the 'Input Product Array' will be considered. The 'Input Product Locations' must have the following columns:
  1. Product (string): The name of each 'Product' in the Market. A WTP Distribution should exist in the 'Input WTP Matrix' for each unique 'Product' name.
  2. Location (string): The point of sale 'Location' of each 'Product' in the Market. A 'Product' can have more than one point of sale 'Location' listed in this table. In addition, the default Location found in the 'Input Product Array' will be added to this list. If the Product is available everywhere to call Customers then this value can be set to 'Global'.
Input Customer Locations: (optional) The ratio of Customer 'Population' at each 'Location'. This table will be ignored if the Input WTP Matrix contains a "Location" column. The 'Input Customer Locations' must have the following columns:
  1. Location (string): The name of each Geographic Location in the Market where Virtual Customers reside.
  2. Population (double): The ratio of Virtual Customers residing at each Geographic Location. If the option 'Locate the Customers' is set to 'In Small Blocks' then these ratios need to be expressed in integers and add to a subset of total 'Number of Customers'.
Input Willingness To Pay Matrix (double): The Willingness To Pay (WTP) Customer Distribution matrix for each Product column in the Market by each Virtual Customer row. The total number of Virtual Available Customers is equal to the number of rows in the WTP Matrix. The 'Input WTP Matrix' must have the following columns:
  1. Product01, Product02, etc (double): Each of the Products listed in the 'Input Product Array' and the 'Input Product Locations' should have a corresponding column in this 'Input WTP Matrix'. Each row represents a different Virtual Customer, and each value represents the Customer's Willingness To Pay (WTP) for each Product.
  2. Location (string - optional): The pre-defined Geographic Location where each Virtual Customer resides. If this Location is missing then it will be automatically allocated according to the 'Input Customer Locations' table.

Output Ports

Output Product Array: The Output Product Array is equivalent to the Input Product Array without any changes. The Product Array is simply passed through the node as a convenience.
Output Willingness To Pay Matrix: The Output Willingness To Pay (WTP) Customer Distribution Matrix corresponds to the Input WTP Matrix. Certain WTP Values within the WTP Matrix will have been shredded where the Product was not available at the Customer's Location. Shredded values are replaced with '0.0'. The Output WTP Matrix is extended to contain these additional columns:
  1. Location: The Geographic Location where each Virtual Customer resides. If this Location was not pre-defined in the 'Input WTP Matrix' then it will be allocated according to the 'Input Customer Locations' table.



To use this node in KNIME, install Market Simulation nodes by Scientific Strategy for KNIME - Community Edition from the following update site:


You don't know what to do with this link? Read our NodePit Product and Node Installation Guide that explains you in detail how to install nodes to your KNIME Analytics Platform.

Wait a sec! You want to explore and install nodes even faster? We highly recommend our NodePit for KNIME extension for your KNIME Analytics Platform. Browse NodePit from within KNIME, install nodes with just one click and share your workflows with NodePit Space.


You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.