Reproduce

The GitHub repository contains the R workflow, dataset documentation, selected outputs, and fitted mixture-model artefacts used by BEAR.

Main workflow

• main.R is the high-level entry point for the repository workflow.
• process/ contains dataset-specific processing scripts.
• workflow/build_bear.R stitches processed inputs into BEAR.rds.
• workflow/fit_mixtures.R fits the mixture models used in downstream summaries.
• workflow/calculate_psr.R calculates the probability of significance, replication, and sign summaries from fitted models, then refreshes the mixture plots for each dataset used by this website.

Data layers

The repository contains code, documentation, and selected outputs. Large combined data artefacts are distributed through GitHub Releases. Individual processed datasets can be fetched through the data submodule for work on individual datasets.

Rerunning BEAR

To work only with the combined dataset, download BEAR.rds from GitHub Releases:

curl -L -o BEAR.rds https://github.com/wwiecek/BEAR/releases/download/v2/BEAR.rds

To work with individual processed datasets, initialise the data submodule:

git submodule update --init --recursive

This downloads the data/ contents needed for work on individual datasets. Dataset-specific processing scripts live in process/. Once processed datasets are present, workflow/build_bear.R builds BEAR.rds; workflow/fit_mixtures.R fits mixture models in mixtures/; and workflow/calculate_psr.R refreshes the PSR table and website mixture plots.