Japanese Personal Name Dataset

A comprehensive dataset of Japanese personal names (first names and last names) with hiragana readings, romaji (Hepburn romanization), and kanji variations.

日本語README

Features

• 5,578 male first names (702 optimized/popular names)
• 3,294 female first names (241 optimized/popular names)
• 2,000 last names with estimated population data
• Multiple kanji variations for each reading
• Romaji (Hepburn) transliterations
• Easy-to-use Python API

Installation

pip install japanese-personal-name-dataset

Or install from source:

git clone https://github.com/shuheilocale/japanese-personal-name-dataset.git
cd japanese-personal-name-dataset
pip install -e .

Dataset Structure

The dataset consists of the following CSV files:

1. first_name_man_org.csv - Male first names (original)
2. first_name_man_opti.csv - Male first names (optimized/popular)
3. first_name_woman_org.csv - Female first names (original)
4. first_name_woman_opti.csv - Female first names (optimized/popular)
5. last_name_org.csv - Last names

Optimized datasets contain curated popular names only.

CSV Format

First Names

Each row represents one name reading:

• Column 1: Hiragana reading
• Column 2: Romaji (Hepburn)
• Column 3+: Kanji variations (variable number)

Example:

あい,ai,藍,愛,亜衣

Last Names

Each row represents one last name:

• Column 1: Kanji
• Column 2: Estimated population
• Column 3: Hiragana reading
• Column 4: Romaji (Hepburn)

Example:

佐藤,1887000,さとう,satou

Usage

Basic Usage

from japanese_personal_name_dataset import load_dataset

# Load the dataset (default: full version)
man_names, woman_names = load_dataset()

# Access male names
print(man_names['たろう'])
# Output: {'en': 'tarou', 'kanji': ['多朗', '多郎', '太朗', '太郎', '大郎']}

# Access female names
print(woman_names['はなこ'])
# Output: {'en': 'hanako', 'kanji': ['花子', '華子', ...]}

Load Optimized Dataset (Popular Names Only)

# Load only popular names
man_names, woman_names = load_dataset(kind='opti')
print(f"Male names: {len(man_names)} types")    # 702 types
print(f"Female names: {len(woman_names)} types")  # 241 types

Include Last Names

# Load with last names
man_names, woman_names, last_names = load_dataset(include_last_names=True)

# Access last name data
print(last_names['佐藤'])
# Output: {'reading': 'さとう', 'en': 'satou', 'count': 1887000}

Using Utility Functions

from japanese_personal_name_dataset import (
    generate_random_name,
    generate_random_full_name,
    search_by_reading,
    search_by_kanji,
    get_last_names,
    is_valid_name,
)

# Generate random name
name = generate_random_name(gender='male')
print(name)  # Example: Taro

# Generate random full name with reading
full_name, reading = generate_random_full_name(gender='female', return_reading=True)
print(f"{full_name} ({reading})")  # Example: Sato Hanako (satou hanako)

# Search by reading (partial match / LIKE search)
results = search_by_reading('kou', partial=True, gender='male')
for r in results[:3]:
    print(f"{r['reading']} ({r['romaji']}): {', '.join(r['kanji'][:3])}")
# Example: kouji (kouji): Koji, Takaji, Yukiharu

# Search by kanji (names containing '子')
results = search_by_kanji('子', partial=True, gender='female')
print(f"Names containing '子': {len(results)} results")

# Get top 10 most common last names
top_10 = get_last_names(limit=10)
for i, name in enumerate(top_10, 1):
    print(f"{i}. {name['kanji']} ({name['reading']}) - {name['count']:,} people")

# Validate name
if is_valid_name('太郎', 'たろう'):
    print("太郎 (tarou) is a valid combination")

Use Cases

• Test data generation for web applications
• Name validation and normalization
• Japanese language learning tools
• Data science and statistical analysis
• Game development (character name generation)

Dataset Statistics

Number of Names

Kanji Variations (per reading)

For original datasets:

• Male names: avg 10 variations, max 447
• Female names: avg 11 variations, max 398

For optimized datasets:

• Male names: avg 45 variations, max 447
• Female names: avg 51 variations, max 291

Data Format

• File format: CSV
• Character encoding: UTF-8
• Line endings: LF
• Romaji system: Hepburn romanization

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

References

• 名字由来net (Myoji Yurai Net)

Disclaimer

While we strive for accuracy, there may be errors in the romanization or kanji variations. This dataset is provided as-is for informational purposes.