PersonaBase: Computational Resources for Data-Driven Personas

Purpose

This repository provides computational resources for data-driven persona research and practice. It contains links to existing datasets, a collection of prompts when using Generative AI in persona development, and example notebooks illustrating how conventional ML algorithms can be used in data-driven persona development.

The purpose of the repository is to advance the scientific study of data-driven personas, particularly by advocating resource sharing and joint benchmarking of results.

NOTE: The code has been designed to run on Google Colab.

How to learn more?

This is work by the Persona Team (https://personateam.xyz). If you are interested in doing a PhD on data-driven personas or other form of research collaboration, reach out to us!

File naming

Taxonomy:

• PD = Persona Development Task Resource (under 'Notebooks')
• PE = Persona Evaluation Task Resource (under 'Notebooks')
• DS = Persona Dataset Resource
• PP = Persona Prompt Resource
• PS = Persona System Resource
• PR = Persona Repository Resource

And:

• a = simulated data
• b = real data (anonym)

So, "PD01a" indicates a persona development task resource with ID of 01 that uses simulated data.

Disclaimer

A major portion of the code has been generated using AI (Claude 4 Sonnet). The code has been verified and tested by humans.

Key contributors

• Joni Salminen (jonisalm@uwasa.fi), University of Vaasa, Finland
• Danial Amin, University of Vaasa, Finland
• Ilkka Kaate, University of Turku, Finland
• Bernard J. Jansen, Qatar Computing Research Institute, Hamad Bin Khalifa University, Qatar

Citation

If you found PersonaBase useful, please use the following citation:

@misc{salminen2025PersonaBase,
      title={PersonaBase: Developing Computational Resources for the Scientific Benchmarking of Data-Driven Personas}, 
      author={Joni Salminen and Danial Amin and Ilkka Kaate and Bernard J. Jansen},
      year={2025}
}