Per-record-type ARSAU analysis pipelines (R-side mirror of morie.arsau_analyze).
Source: R/arsau_analyze.R
arsau_analyze.RdEach public callable in this file loads one ARSAU dataset via the
morie_arsau_load_* loaders defined in R/arsau.R and
chains the jurisdiction-agnostic MRM Use-of-Force primitives from
R/mrm_uof.R over it, returning a single named-list result
(classed c("morie_arsau_result", "morie_rich_result", "list"))
that bundles the loaded data, every sub-analysis, and a
multi-paragraph natural-language interpretation.
Details
These analyzers do NOT invent new statistical methods. They wire
the generic mrm_uof_* callables against the column names
that the Ontario open-data release actually publishes. If the
upstream schema changes, the generic callables in
R/mrm_uof.R continue to work; only the column-name constants
below need patching.
Public callables:
morie_arsau_analyze_main_records(per year, 2023 / 2024)morie_arsau_analyze_individual_records(per year, 2023 / 2024)morie_arsau_analyze_probe_cycle_records(per year, 2023 / 2024)morie_arsau_analyze_weapon_records(per year, 2023 / 2024; 2023 requiresallow_invalid = TRUE)morie_arsau_analyze_aggregate_summary(2020-2022)morie_arsau_analyze_detailed_dataset(2020-2022)
Every analyzer returns a list whose named slots (data,
sidecar, force_concentration, data_quality,
disparity_by_race, ...) hold the constituent sub-results,
so callers can drill into individual tests without re-running the
full pipeline.
References
Ontario Ministry of the Solicitor General. Annual Report on Special and Adaptive Units / Data on Police Use of Force in Ontario: 2020-2022, 2023, and 2024 releases. https://data.ontario.ca/dataset/police-use-of-force-race-based-data. Technical notes accompanying each annual release describe the data-quality reasons for the 2023 weapon-records invalidity flag.