Generate Realistic Persons#

Data Designer’s sampler column type can be used to sample realistic person data and synthetic personas. Generated using Data Designer itself, as well as a Probabilistic Graphical Model trained on census data, the sampled datasets are grounded in real-world demographic, geographic and personality trait distributions to capture the diversity and richness of the population.

Person Objects in Data Designer#

Creating Person Samplers#

Person samplers generate realistic person entities with configurable attributes and optional synthetic persona data. Each sampler creates a different person object that you can reference throughout your data design.

builder.add_column(
    name="customer",
    column_type="sampler",
    sampler_type="person",
    params={
        "locale": "en_US",
        "sex": "Male",
        "with_synthetic_personas": True
    },
)

builder.add_column(
    name="employee",
    column_type="sampler",
    sampler_type="person",
    params={
        "locale": "en_GB",
        "sex": "Female",
        "with_synthetic_personas": False
    },
)

builder.add_column(
    name="random_person",
    column_type="sampler",
    sampler_type="person",
)

Configuration Options#

Person samplers accept these configuration parameters:

Basic Configuration:

sex: Specify “Male” or “Female” (optional)
locale: Language and region code (optional, e.g., “en_US”, “fr_FR”, “de_DE”)
city: Filter on cities within the specified locale (optional)
age_range: Age range for filtering (default: ages above 18 only)
state: Filter on US states, only valid when locale is set to “en_US” (optional)

Synthetic Personas Configuration:

with_synthetic_personas (default: False): When set to True, samples detailed personality profiles, cultural backgrounds, skills, interests, and context-specific personas for comprehensive character modeling. The personas are sampled from expanded versions of NVIDIA’s Nemotron-Personas datasets.

Filtering Notes:

When using US locale (“en_US”), you can filter on age range, sex, city, and state
For non-US locales, filtering is limited to age range, sex, and city only
You can choose either city or state when filtering, not both

Locale Support and Data Quality#

Grounded in real-world demographic data: The best quality person / synthetic persona data is generated by sampling a superset of the Nemotron-Personas datasets. This is currently supported for the following locales: “en_US”, “ja_JP”, “hi_IN”, and “en_IN”.

Other Locales: Data Designer uses the Faker library to generate person data (synthetic personas are not supported). While Faker provides basic attributes like names and addresses, it doesn’t maintain the same demographic accuracy or attribute relationships as the Nemotron-Personas datasets.

Synthetic Personas: Available for all locales when with_synthetic_personas=True. Persona generation adapts to cultural context based on the specified locale and demographic information.

Person Data Structure#

Core Demographic Fields (Always Available)#

Field Name	Type	Description
uuid	str	Unique identifier
first_name	str	Person’s first name
last_name	str	Person’s last name
sex	categorical	Person’s sex (Male or Female)
age	int	Person’s age
street_number	int	str
street_name	str	Name of the street
city	str	City name
country	str	Country name
marital_status	categorical	None
education_level	categorical	None
bachelors_field	categorical	None
occupation	str	None
birth_date	date	Calculated birth date based on age
locale	str	Locale

In addition to the above fields, person objects also contain locale-specific fields for such as “state” for “en_US” or “area” for “ja_JP”.

Personality Traits (Available when `with_synthetic_personas=True`)#

Big Five personality model with t-scores and interpretive labels:

Field Name	Type	Description
openness	dict	Openness to experience (t_score, label, description)
conscientiousness	dict	Conscientiousness (t_score, label, description)
extraversion	dict	Extraversion (t_score, label, description)
agreeableness	dict	Agreeableness (t_score, label, description)
neuroticism	dict	Neuroticism (t_score, label, description)

Each personality trait contains:

t_score: Standardized score (typically 0-100)
label: Interpretive label (“low”, “average”, “high”, “very high”)
description: Detailed behavioral description

Synthetic Persona Fields (Available when `with_synthetic_personas=True`)#

Background and Development#

Field Name	Type	Description
cultural_background	str	Detailed narrative about cultural influences and upbringing
skills_and_expertise	str	Comprehensive description of professional and personal capabilities
skills_and_expertise_list	str	List format of key skills and competencies
hobbies_and_interests	str	Detailed description of personal interests and activities
hobbies_and_interests_list	str	List format of hobbies and interests
career_goals_and_ambitions	str	Professional aspirations and long-term objectives

Persona Profiles#

Field Name	Type	Description
persona	str	Brief summary personality profile
detailed_persona	str	Comprehensive personality and behavioral description
professional_persona	str	Work environment personality and career approach
finance_persona	str	Financial decision-making style and money management approach
healthcare_persona	str	Health and wellness attitudes and behaviors
sports_persona	str	Sports interests and physical activity preferences
arts_persona	str	Artistic tastes, cultural interests, and creative preferences
travel_persona	str	Travel style, preferences, and exploration approach
culinary_persona	str	Food interests, cooking style, and dining preferences

Best Practices#

Choosing Configuration Options#

Use locales that are backed by a Nemotron-Personas dataset for maximum demographic accuracy and realism
Enable with_synthetic_personas=True when you need rich character development, personalized content generation, or comprehensive behavioral modeling
Disable synthetic personas for basic demographic testing or when computational efficiency is prioritized

Effective Persona Usage#

Match persona depth to use case: Use basic personas for simple applications, detailed personas for comprehensive character modeling
Leverage context-specific personas: Use professional_persona for workplace scenarios, culinary_persona for food-related applications
Combine multiple persona fields in prompts for richer, more nuanced content generation

Performance Considerations#

Synthetic personas add processing time: Only enable when the additional data provides value
Cache person objects when using the same personas across multiple columns
Consider batch generation for large datasets requiring consistent persona quality

Quality Assurance#

Validate persona consistency: Ensure generated content aligns with personality traits and demographic information
Test across different locales to understand quality variations
Review persona coherence when using multiple context-specific personas for the same individual