Make your characters more representative and realistic.
A REST API and command-line tool for probabilistically generating random character profiles from a given input location using real-world demographic data. Generating a new persona rolls the dice on features such as age, sex, sexuality, ethnicity, language and religion. This project was born out of a lack of tools for building representative and realistic characters for stories.
https://persona-api.vercel.app/v1/<location>/
$ curl https://persona-api.vercel.app/v1/england/
[
{
"age": 21,
"sex": "Female",
"sexuality": "Heterosexual",
"ethnicity": "British, White",
"religion": "Christian",
"language": "English",
"location": "Oldham, North West"
}
]
Multiple personas from the same location can be generated at once by providing a count
query parameter.
https://persona-api.vercel.app/v1/<location>/?count=5
All locations currently included can be listed with the /v1/locations/
endpoint.
https://persona-api.vercel.app/v1/locations/
$ curl https://persona-api.vercel.app/v1/locations/
[
"australia",
"canada",
"germany",
"global",
"united_kingdom",
"england",
"london",
"northern_ireland",
"scotland",
"wales",
"california",
"florida",
"texas"
]
Currently, not all features are available for each location. For a given location, all features available for generation can be retrieved with the /v1/<location>/features/
endpoint.
https://persona-api.vercel.app/v1/<location>/features/
$ curl https://persona-api.vercel.app/v1/england/features/
{
"england": [
"age",
"sex",
"sexuality",
"religion",
"ethnicity",
"language",
"location"
]
}
Install Python dependencies from requirements.txt
.
pip install -r requirements.txt
Run main.py
from the root directory.
python src/main.py <location>
The generated persona can be limited to specific features using the feature flags to include.
python src/main.py <location> --age --location --language
Multiple personas can be generated at once using the -n
flag.
python src/main.py <location> -n <count>
python src/main.py united_kingdom
> United Kingdom
Age: 48
Sex: Female
Sexuality: Heterosexual
Ethnicity: British, White
Religion: No religion
Language: English
Location: Blackburn with Darwen, North West, England
The demographic data is carefully sourced from reputable census data for each location. Sources for each location can be found alongside the data in each README.md
in /data
.
The full list of locations currently available can be found here. It includes countries, groups of locations (e.g. UK, USA), and cities. More locations and features will continue to be added in future.
Personas generated are basic approximations. Character features are naively generated under the assumption that each feature is independent from one another. This assumption is not true; knowing a person's age could help you better predict their religion. However, the sourcing of accurate and large scale data necessary for the joint probabilities for all feature combinations is exponentially harder to achieve. As a result, generated characters should be taken with a pinch of salt, and very occasionally personas will be generated that have a combination of features that may seem extremely unlikely or even impossible. Obviously, the fewer features included in the persona, the easier it is to approximate, and the less likely this is to occur.
Contributions are very welcome for data or general improvements.
To contribute:
- Fork the repo.
- Create your feature branch (
git checkout -b my-new-feature
) - Commit your changes (
git commit -am "Add some feature"
) - Push to the branch (
git push origin my-new-feature
) - Create a new pull request
When contributing data, keep content, directory structure and JSON formatting consistent and remember to note your source (including URL) in data/.../<location>/README.md
. Sources should be from reputable organisations conducting census research. Avoid "Other" as a feature attribute. Do not worry if percentages do not sum to 1 exactly, all feature probabilities are normalised during generation.