Skip to contents

Generic Multilevel Reconciliation Methodology (MRM) Use-of-Force callables

Source: R/mrm_uof.R
mrm_uof.Rd

Six jurisdiction-agnostic analyses for police Use-of-Force data, mirroring the Python module morie.mrm_uof. Every function accepts a data.frame (or tibble) and returns a named list carrying both the numeric outputs and a multi-paragraph plain-language interpretation, so the result can be printed to a notebook without further post-processing.

Details

Functions

  • mrm_uof_force_concentration: Hill-MLE Pareto exponent + Gini coefficient + top-5 / top-10 share for incident counts aggregated by force / service.

  • mrm_uof_weapon_diversity: weapon-by-force contingency: chi-square, Cramer's V, and the top-3 cells by standardised Pearson residual.

  • mrm_uof_yoy_change: year-on-year percentage change with a manual largest-gap change-point fallback (the R side does not require ruptures).

  • mrm_uof_region_locality: region-at-time vs. region-now contingency: diagonal share, chi-square, Cramer's V.

  • mrm_uof_demographic_disparity: per-category outcome rates with Wilson 95\ baseline group, optional non-parametric bootstrap percentile interval on the risk ratio.

  • mrm_uof_data_quality_audit: per-column null and dtype audit, with optional schema-comparison against a supplied CKAN sidecar list or column-spec list.