Guarded Evidence Synthesis for the Indian Microplastics Evidence Database

A research-grade, guardrail-first interface to the Indian Microplastics Evidence Database (IMED). It provides release-data access, relational and scientific validation, comparability-aware filtering, descriptive and inferential statistics, evidence-synthesis eligibility, meta-analysis helpers, publication figures, mapping, provenance and citation retrieval, and manuscript-oriented outputs. Unsafe pooling across incompatible matrices, denominators, mass bases and comparability classes is blocked by default.

imedR 0.3.3

imedR is a manuscript-grade, guardrail-first R package for the Indian Microplastics Evidence Database (IMED).

Frozen v1.0 snapshot represented in this build

Studies: 97
Sites: 217
Samples: 293
Measurements: 313
Comparability classes: 17

The bundled analytical snapshot is generated directly from IMED_v1.0_FROZEN.xlsx; workflow and extraction-log sheets are excluded.

Capabilities

relational/scientific validation;
measurement enrichment with IMED comparability metadata;
guarded analysis-set construction;
explicit HOLD handling;
descriptive and conservative inferential statistics;
meta-analysis eligibility checks;
guarded random-effects raw-mean and prevalence meta-analysis helpers;
adversarial guardrail tests and known-answer statistical tests;
size-window diagnostics and evidence-synthesis sensitivity summaries;
leave-one-study-out influence analysis and guarded meta-regression;
forest/funnel diagnostics and a consolidated validation report;
publication-oriented abundance, composition and QA/QC figures;
coordinate-class-aware mapping;
missingness and evidence-profile diagnostics;
primary citation and field-level provenance retrieval;
manuscript-oriented tables and result tidying;
high-resolution figure export.

Core rule

Validate -> define a scientifically compatible analysis set -> analyse -> visualize -> retrieve provenance/citations.

The package is designed to stop analyses that look computationally possible but are scientifically indefensible.

Validation status

The test suite passes (27 tests, 0 failures, 0 warnings, 0 skipped), and the IMED v1.0 relational integrity checks report no ERROR-level failures. R CMD check --as-cran was run on the Windows build service (win-builder, R 4.6.1, 8 October 2026) for version 0.3.1 and returned no errors and no warnings; the only note was the standard "New submission" note for a package not yet on CRAN.

Release-normalization note (v0.3.1)

The bundled frozen IMED v1.0 source snapshot is preserved. At load time, imedR removes wholly empty export rows and disambiguates the second genuine ST0064 size-method record from SZ0064-01 to runtime ID SZ0064-02. This is an identifier-only normalization; no scientific value is altered.

Installation

# install.packages("remotes")
remotes::install_github("ChanikyaNaiduCN7/imedR")
library(imedR)
imed_version()

Optional packages for mapping, meta-analysis and tables (sf, metafor, gt) are installed only if you use those functions.

Citation

Naidu C (2026). imedR: Guarded Evidence Synthesis for the Indian Microplastics Evidence Database. R package version 0.3.3. https://doi.org/10.5281/zenodo.23236915

Please also cite IMED/IMEA v1.0 (https://doi.org/10.5281/zenodo.23180568) and the original primary publications for any study-specific claim.

License

Package code: MIT (see LICENSE). The IMED/IMEA data release is licensed separately (see LICENSE_AND_REUSE.md in the data release). Third-party source rights remain with the original rights holders.