Software

Select works

Reproducible analysis pipelines

• targets, dynamic function-oriented Make-like workflows for R. Long-term successor to drake.
• tarchetypes, general-use archetypes for targets and pipelines, including parameterized Quarto at scale.

Bayesian statistics pipelines

• stantargets, Reproducible Bayesian data analysis pipelines with targets and cmdstanr.
• jagstargets, Reproducible Bayesian data analysis pipelines with targets and R2jags.

Bayesian methods for historical borrowing in clinical trials

• historicalborrow: non-longitudinal hierarchical and mixture models for historical borrowing.
• historicalborrowlong: a longitudinal hierarchical model for historical borrowing.

Distributed computing

• crew: a mirai-powered worker launcher framework.
• crew.cluster: crew launcher plugins for traditional clusters.
• crew.aws.batch: a crew launcher plugin for AWS Batch.

Other packages

R package utilities and infrastructure

• proffer A pprof-based profiler for R code.
• txtq A small message queue for parallel processes.
• downsize Ease the transition between testing and deploying data analysis workflows.

Dissertation research

• fbseq Fully-Bayesian GPU-accelerated analysis of RNA-seq data (arXiv preprint).
• fbseqCUDA CUDA GPU support for fbseq.
• fbseqOpenMP OpenMP support for fbseq.

Prior work on reproducible analysis pipelines

• gittargets, local data version control for the targets R package.
• drake an R-focused pipeline toolkit for data analysis projects. Featured in the rOpenSci blog and R Podcast Episode 22.
• remakeGenerator Generate large remake-style workflows.
• parallelRemake Parallel computing support for remake.