Research Software

I build research software so other people can run my methods themselves, drawing on more than a decade of experience in statistics and industry.

On GitHub

My open-source work, packages, teaching material, and reproducible analyses are on GitHub.

Alyssa Columbus GitHub contribution graph

Maintained Software

R
Specify, run, and report vibration-of-effects and specification analyses. Turns a set of defensible analytic choices into a reproducible grid of results, with reports that show how widely the estimates vary. Ships with a companion Quarto book.
author and maintainer
R
A client for Amazon Simple Storage Service (S3), used across the R ecosystem for cloud data access, and one of the R packages I’ve authored or maintained that together have been downloaded more than 50 million times from the Comprehensive R Archive Network (CRAN).
contributor and former maintainer
R
Interactive applications for teaching statistical thinking and for exploring analytic choices, used in workshops and in the classroom.

Software Peer Review

I review research software submissions for the Journal of Open Source Software, rOpenSci, and pyOpenSci, assessing correctness, documentation, test coverage, and whether a package will still work for someone in three years.

  • Software referee, the Journal of Open Source Software, rOpenSci, and pyOpenSci.
  • Journal referee, Journal of Machine Learning Research, Transactions on Machine Learning Research, and The R Journal.
  • Instructor and lesson maintainer, The Carpentries.

Industry Practice

I spent about six years working with data in industry before starting my doctorate, as a data scientist, an information security analyst, and a consultant. That experience is why securing the agents that run analyses is one of my research threads: I’ve been on the side of the table where a pipeline has to keep running, a dataset has to stay protected, and a decision has to be defended to someone who didn’t write the code.

  • Artificial intelligence (AI) safety, policy, and governance: independent advisory work for Google DeepMind, Anthropic, and OpenAI on the safety and oversight of frontier AI systems.
  • Data science: statistical modeling, pipeline engineering, and analysis in production, including breast cancer risk prediction with the Athena Breast Health Network as a data scientist at Pacific Life, and selection to the McKinsey & Company Insight program (one of roughly 100 from a field of more than 2,000).
  • Consulting and strategy: a strategy and analytics consultant at Deloitte, translating statistical findings for the people who have to act on them.
  • Information security and privacy: security analysis and data governance, with the Global Information Assurance Certification Security Essentials (GSEC) credential and the OneTrust Fellow of Privacy Technology designation.
Working Together

I take on selected consulting and advisory work in study design, analytic reproducibility, and the security of AI systems used on sensitive data. Get in Touch.

Open Source & Community

I’ve spent years building communities that make open-source statistical computing welcoming to newcomers, organizing more than 70 meetups for a combined membership of more than 2,000 statisticians, data scientists, and students, who rated them 4.8 out of 5. That work runs from founding the Irvine chapter of R-Ladies Global, which grew past 600 people, to co-organizing its Baltimore group, now more than 1,400 strong, and a term on the organization’s international leadership team.

Group photo of members at the launch meetup of R-Ladies Irvine, in front of a slide reading Launching R-Ladies Irvine
Launch meetup of R-Ladies Irvine
Alyssa presenting at an R-Ladies Irvine meetup next to a slide titled Find us online
Presenting at an R-Ladies Irvine meetup