KNIME Nodes
Advanced Data Generation
In most of the workflows found in this blog, the input data was generated by a Market Simulation node, such as:
- Customer Distributions node
- Matrix Distributions node
But KNIME itself offers many powerful data generation capabilities – particularly in the KNIME Data Generation Extension. To install this extension, check out the Getting Started Guide: Installing the KNIME Analytics Platform.
This workflow generates Customer Profile sample data using some of these advanced data generation techniques. All of the nodes used in this example come from KNIME. But these nodes and this type of data generation can be easily integrated into a Market Simulation workflow.
This KNIME Node Use Case provides an example of a useful KNIME workflow. These workflows do not depend upon Market Simulation but can supplement a Market Simulation workflow. If you have not yet installed KNIME, go to Getting Started.
Downloads
#1 New Customers
This first step will create 200 random Customers (48% Male / 52% Female) with Customer RowID’s running from c0 to c199 and CustomerID’s running from 1 to 200.
Empty Table Creator
Create a column of 200 Customers in an Empty Table.
Column Rename
Rename the CustomerID Column.
Counter Generator
Generate a Counter for each CustomerID.
Random Label Assigner
Randomly assign Customers to be either Male (48%) or Female (52%).
Output Data
Adds the CustomerID and the Gender data columns.
#2 Age Profile
This second step uses 6 Age Pyramid Profiles to randomly allocate the age of these Customers from between 17 years old and 100 years old.
Random Label Assigner
Create a set of Population Age Pyramids.
Double To Int
Round the generated Age.
Gaussian Distributed Assigner
Define the probable Age for each Pyramid.
Column Filter
Remove the Pyramid Profile column.
Output Data
Adds the Age data column.
#3 Occupations
Assign an Occupation to each Customer based upon their Age and an Occupation Probability.
Numeric Binner
Bin Customers by Age into Generations.
Conditional Label Assigner
Assign a Family status to each Customer by their Age Generation.
Gaussian Distributed Assigner
Assign Occupations to each Customer randomly by their Age Generation.
Output Data
Adds the Generation, Occupation, and Family columns.
#4 Income
Generate an appropriate Income for all Customers based upon their Age Generation and whether or not they are a Student.
Nominal Value Row Filter
Filter by Occupation = Non-Student.
Nominal Value Row Filter
Filter by Occupation = Student.
Gamma Distributed Assigner
Generate Incomes by Age Generation.
Gamma Distributed Assigner
Generate Incomes randomly for Students.
Output Data
Adds the Income column.
#5 Shred Income
Shred the Income (that is, set Income = 0.0) of some Customers based upon the Probability that they are not currently working.
Conditional Label Assigner
Shred Income by the Probability the Customer is working.
Constant Value Column
Shred Income by overriding Income column and setting to 0.0.
Row Splitter
Split rows so Top = Has Income and Bottom = No Income.
Output Data
Deletes some of the Income data.
#6 Clean Up
Clean up the final Customer Profiles by rounding the Income, removing extra columns, and sorting by CustomerID.
Double To Int
Round the Income to the nearest Integer.
Sorter
Sort all the rows by the original CustomerID.
Column Filter
Remove the Generation column and the Has Income column.
Output Data
Removes the Generation and Has Income columns then sorts.
