Numeric Transformers

Transform and mask numeric values

Numeric Transformers

Transform numeric values while preserving statistical properties or adding privacy protection.

add_noise_percentage

Add random noise as a percentage of the original value.

Category: Numeric Type: Transformer Data Types: integer, decimal, float

Parameters

Parameter	Type	Default	Description
`noise_percent`	float	10	Percentage of noise (±%)

Examples

10% noise (default):

columns:
  - name: salary
    function: add_noise_percentage

Small noise (5%):

columns:
  - name: price
    function: add_noise_percentage
    parameters:
      noise_percent: 5

Large noise (25%):

columns:
  - name: revenue
    function: add_noise_percentage
    parameters:
      noise_percent: 25

Behavior

Input: 100
With noise_percent: 10
Output: Random value between 90 and 110

Use Cases

Protect exact financial figures
Maintain statistical distributions
Allow trend analysis while obscuring specifics

round_to_range

Round numbers to ranges for privacy.

Category: Numeric Type: Transformer Data Types: integer, decimal, float, string

Parameters

Parameter	Type	Default	Description
`range_size`	float	10	Size of range buckets

Examples

Age ranges (10-year buckets):

columns:
  - name: age
    function: round_to_range
    parameters:
      range_size: 10

Salary bands ($10,000):

columns:
  - name: salary
    function: round_to_range
    parameters:
      range_size: 10000

Behavior

Input: 47
With range_size: 10
Output: "40-50"
Input: 85000
With range_size: 10000
Output: "80000-90000"

Use Cases

Age bracketing for analysis
Salary band reporting
Privacy-preserving analytics

Comparison

Function	Output Type	Preserves	Use Case
`add_noise_percentage`	Number	Statistical distribution	Financial analysis
`round_to_range`	String (range)	General magnitude	Demographic data

Common Patterns

Financial Data

columns:
  - name: account_balance
    function: add_noise_percentage
    parameters:
      noise_percent: 5
  - name: transaction_amount
    function: add_noise_percentage
    parameters:
      noise_percent: 10

Demographics

columns:
  - name: age
    function: round_to_range
    parameters:
      range_size: 5  # 5-year age groups
  - name: income
    function: round_to_range
    parameters:
      range_size: 25000  # Income brackets

Healthcare

columns:
  - name: weight_kg
    function: add_noise_percentage
    parameters:
      noise_percent: 3
  - name: height_cm
    function: add_noise_percentage
    parameters:
      noise_percent: 2

Tips

Choosing Noise Level

Data Type	Suggested Noise
Precise measurements	2-5%
Financial amounts	5-15%
Sensitive personal data	15-25%

Preserving Aggregates

Low noise percentages (3-5%) maintain aggregate statistics while obscuring individual values.

Range Size Selection

Age: 5 or 10 years
Salary: $10,000 or $25,000
Revenue: Based on business scale

Number Generators - Generate new numbers
Date Transformers - Transform dates
Masking Functions - General masking

Numeric Transformers