# Trace data pipeline for open-source coordination research

> 8 open-source projects, 1,169,489 actions, 93 major releases, prepared as time-ordered event sequences. Published in MIS Quarterly (2024).

- Role: Data infrastructure, co-author
- Period: 2017 to 2023, Cologne Institute for Information Systems
- Status: Published
- Stack: GitHub API, Digital trace data, Relational event modelling, Data reconciliation
- Paper: https://doi.org/10.25300/MISQ/2023/16789

## Context and problem

Open-source projects coordinate through commits, issues and releases. These trace records show what happened, not why. The study needed clean, time-ordered event sequences from many repositories before any model could run.

## Approach

- Pulled commit and release data from the GitHub platform API.
- Reconciled actors and events so they could be ordered in time.
- Loaded the result into event sequences for relational event modelling.
- The research team chose relational event modelling early in the project. I designed, built and ran the data collection and preparation pipeline, and worked closely with the team on interpreting what the results said about developer behaviour.

## Result

The study, *The Entrainment of Task Allocation and Release Cycles in Open Source Software Development*, appeared in MIS Quarterly in 2024. I am fourth of five authors.

## Transfer

Major releases are events around which open-source communities self-organise. The same pipeline could be re-run as planned work, once algorithmic contributors such as bots and AI reviewers enter the population.