No packages match

recoverlite - Pre-Data Recovery Tests for Planned Study Designs

A prototype implementation of the pre-data recovery test: a standardized simulation protocol for evaluating whether a planned design-analysis pair can recover its target estimand before data collection. Researchers declare the estimand, data-generating assumptions, data strategy, missing-data process, and analysis strategy; the package simulates a crossed scenario grid (null and target effects, each under declared and pessimistically perturbed nuisance assumptions), applies the planned analysis to each simulated dataset, and reports target-null rejection, power, target bias with its exact decomposition into estimator bias and estimand drift, coverage, Type S and Type M errors, precision, and classified model failure, each with Monte Carlo uncertainty and explicit inclusion rules. A versioned threshold profile converts the diagnosands into a PASS/RISK/FAIL verdict - a decision convention, not a validity classification. Built on the 'DeclareDesign' framework.

Last updated

clinical-trialsdeclaredesignestimandsmetasciencemonte-carlopower-analysispreregistrationresearch-methodssimulationstatisticsstudy-design

3.00 score 1 stars 1 scripts

assesslite - AssessLite: Test What Your Analysis Depends On

Structural assumption assessment for causal analysis. Makes the invariance assumptions behind a result explicit, attacks them with transformation tests (unit permutation, cluster holdout, temporal split, subgroup stability), returns three-way verdicts (stable, unstable, not resolvable), and exports the reasoning path as a durable audit record. AssessLite is the product, structural_audit() runs the assessment, and write_audit() writes the audit record. Sibling to 'recoverlite': recoverlite asks whether a method can recover the quantity it estimates; AssessLite asks what a result depends on and whether it survives the violation of those assumptions. R implementation of the AssessLite core specification v0.1. Records claim-to-source support, never truth: a stable verdict means the conclusion survived the attacks that were run, at the precision the data allowed.

Last updated

python

3.00 score 1 stars

vahtian - Reproducible, Provenance-First Evidence Tooling

Freeze a record set into a content-hashed, provenance-stamped, date-locked corpus; verify reproducibility; and keep a hash-chained, tamper-evident audit trail. The canonical on-disk format and the content hash are byte-identical with the Python package 'vahtian', so frozen corpora and audit ledgers interoperate across the two languages.

Last updated

citationsevidence-synthesislocal-firstresearchresearch-integrityresearch-methodssystematic-reviewzotero

2.70 score 1 stars 1 scripts

qualivahti - Local-First Qualitative Analysis Scripts (QualiVahti)

Runners for the QualiVahti analysis scripts: coding review summaries, code frequency tables and charts, code-by-participant and theme-by-participant matrices, code co-occurrence network drawings, publication-ready descriptive tables, an AI-use disclosure draft, and a pre-flight defensibility check, built from human-reviewed qualitative coding. All processing is local and only rows a human reviewer marked accepted or edited are counted. The scripts ship inside the package (see qv_scripts_path()) and can also be run directly with Rscript; they are the open analysis engine of the QualiVahti Local vault <https://vahtian.com/qualivahti-local/>.

Last updated

local-firstobsidianpythonqualitative-data-analysisqualitative-researchtranscription

2.48 score 2 stars

vahtian.epinet - R Interface to the EpiNet Toolkit

An R interface to EpiNet, a transparent toolkit for honestly evaluated outcome models on tabular and graph-shaped data. It wraps the tested Python 'vahtian.epinet' package through 'reticulate', so the algorithms are single-sourced and cannot diverge across languages. It exposes the same selectable estimators (random forest, logistic regression, optional XGBoost) and returns calibrated, caveated results: discrimination, calibration, bootstrap intervals, a label-permutation null, and feature importance. A research and education demonstrator, not clinical decision support.

Last updated

calibrationclinical-prediction-modelsdata-scienceepidemiologyfederated-analysismachine-learningnetwork-analysispythonreproducibility

2.48 score 1 stars

palettecore - Derive, Optimise and Audit a Scientific Colour Palette from One Seed Colour

Generates sequential, diverging and categorical colour palettes from a single seed colour in OKLCH (the cylindrical lightness-chroma-hue representation of the Oklab perceptual colour space), with spacing measured by the CIEDE2000 colour-difference formula of the International Commission on Illumination. Audits every palette under simulated colour-vision deficiency, greyscale conversion, the standard Red Green Blue (sRGB) gamut and Web Content Accessibility Guidelines (WCAG) contrast. Colour-vision deficiency is simulated at severity 1.0 with the model of Machado, Oliveira and Fernandes (2009) <doi:10.1109/TVCG.2009.113>; the design rationale follows Crameri, Shephard and Heron (2020) <doi:10.1038/s41467-020-19160-7>. Mirrors the 'Python' reference implementation maintained in the same repository and is validated against shared parity fixtures. Thresholds are configurable design rules, not established accessibility cut-offs.

Last updated

accessibilityciede2000colorcolorblinddata-visualizationoklchpalettepython

2.00 score 1 stars 3 scripts 188 downloads

forskai - Recoverability Testing for Research Design

Public entry point for ForskAI. Recoverability testing examines whether a planned study can recover the effect it is meant to detect, before data collection begins -- design evidence under stated assumptions, not a guarantee of results. The Design Pilot service runs at <https://www.forskai.com>. This package currently carries the project's identity and will grow into the thin client for the hosted engine; the recoverability engine itself is not distributed here.

Last updated

1.70 score 1 stars

palettecore - Derive, Optimise and Audit a Scientific Colour Palette from One Seed Colour

Generates sequential, diverging and categorical colour palettes from a single seed colour in OKLCH (the cylindrical lightness-chroma-hue representation of the Oklab perceptual colour space), with spacing measured by the CIEDE2000 colour-difference formula of the International Commission on Illumination. Audits every palette under simulated colour-vision deficiency, greyscale conversion, the standard Red Green Blue (sRGB) gamut and Web Content Accessibility Guidelines (WCAG) contrast. Colour-vision deficiency is simulated at severity 1.0 with the model of Machado, Oliveira and Fernandes (2009) <doi:10.1109/TVCG.2009.113>; the design rationale follows Crameri, Shephard and Heron (2020) <doi:10.1038/s41467-020-19160-7>. Mirrors the 'Python' reference implementation maintained in the same repository and is validated against shared parity fixtures. Thresholds are configurable design rules, not established accessibility cut-offs.

Last updated

1.00 score 3 scripts 215 downloads