No packages match

rmoriebricklayer - Reproducible Data Capsules with Provenance and Fallback

Tools for building brick-proof, reproducible, self-contained data capsules. Resolves open-data sources through the Comprehensive Knowledge Archive Network ('CKAN', <https://ckan.org/>) package_show and package_search endpoints, records and verifies provenance with Secure Hash Algorithm 256 ('SHA-256') digests and Internet Archive 'Wayback Machine' (<https://web.archive.org/>) snapshots, validates downloaded data against a pinned schema, and falls back to schema-driven synthetic data when the real source is unreachable. Run records are captured in a manifest plus a plain-language summary so any result can be traced back to its inputs. Distributional drift between a pinned capsule and a fresh fetch is tested with Kolmogorov-Smirnov, chi-square, population stability index, Jensen-Shannon divergence and 'Benford' first-digit screens, because a re-released extract can be statistically identical yet differ byte-for-byte, and a column can keep its name and type while having been silently rescaled. Manifests can be authenticated rather than only checksum-verified, with keyed digests ('HMAC-SHA-256', RFC 2104) or post-quantum hash-based signatures ('Winternitz' one-time signatures under a 'Merkle' tree, RFC 8391), and pinned chunk-wise through a 'Merkle' tree so a mismatch identifies which part of a capsule moved. Also ships a compiled C++ core that sibling packages in the 'rmorie' ecosystem reach through 'LinkingTo' for a single, shared numeric and provenance-hashing backend: summary, robust and rank statistics, 'SHA-256', 'SHA-512' and 'CRC-32', and, implemented from scratch with no system 'cryptography' library, the post-quantum standards 'ML-KEM' (FIPS 203), 'ML-DSA' (FIPS 204) and 'SLH-DSA' (FIPS 205), 'HQC' in two revisions, 'XMSS' (RFC 8391), the 'AES-256' 'CTR_DRBG' of SP 800-90A, 'SHA-3' and 'SHAKE' (FIPS 202), 'BLAKE2b', 'PBKDF2', an ASN.1/DER parser with 'ECDSA' (P-256, P-384, P-521) and 'RSA' ('Rivest-Shamir-Adleman') signature verification, and 'X.509' certificate chain, 'OCSP' (Online Certificate Status Protocol) and RFC 3161 timestamp verification. The standardised schemes are checked byte for byte against 'OpenSSL' 3.5 and the known-answer vectors of 'NIST' (the US National Institute of Standards and Technology) in the test suite. Timing is measured on hardware, power leakage is assessed in simulation, and 'ML-KEM decapsulation' and 'ML-DSA' signing are first-order masked; none of this code has had a third-party security audit, and it is offered for provenance and research use, not as a vetted security library. For the published administrative tables these capsules usually hold, it computes period-over-period change matched on the period rather than the row, with the exact conditional-binomial interval for a ratio of counts and with a percentage-point reading kept distinct from a percent change, rendered to Hypertext Markup Language ('HTML'), Portable Document Format ('PDF'), delimited text, JavaScript Object Notation ('JSON') or Markdown. Interval categories such as "2 to 5" or "50+" are parsed to bounds and the dependence of any derived figure on the open top band is measured rather than assumed. Concentration is summarised by the 'Gini' coefficient, the Lorenz curve and tail-index estimation by exact discrete maximum likelihood; trend in a series of a few periods by the Mann-Kendall test with 'Theil-Sen' slopes, a permutation step-change scan and Poisson rate ratios; and region-coded counts by indirect standardisation, exact standardised incidence ratios, the empirical Bayes shrinkage of Clayton and 'Kaldor' (1987) <doi:10.2307/2532003>, funnel-plot limits and Moran's I. Self-exciting event series are fitted as 'Hawkes' processes (exponential, 'Weibull', gamma and 'Lomax' kernels, constant or sinusoidal baseline) by exact maximum likelihood with time-rescaling residuals, and custodial populations are summarised by average daily population, average length of stay and stock-and-flow measures in the formulation of 'Lakner'. Air pollution is followed from source to burden: 'Pasquill-Gifford' dispersion coefficients with the 'Briggs' plume rise, Gaussian plume and puff, Eulerian and Lagrangian models; concentration-response functions for PM2.5 and NO2 (the WHO 2021 guideline review and the 'Burnett' integrated exposure-response curve), the 'Levin' attributable fraction, displaced mortality, a cross-fitted partially linear exposure-response estimator with a bootstrap, and the 'Wagstaff' concentration index of exposure by income; and the electricity and carbon footprint of a computation, measured from the CPU's energy counters when the operating system exposes them or modelled with the 'Green Algorithms' method against a sourced carbon-intensity table.

Last updated

academic-researchagpl-3-0cckancppcriminologycryptographyhawkes-processmorieopen-datareproducibilityrmoriermoriedatasiusocial-sciencesociolegal-studiespecial-investigations-unitcurlzlibcpp

7.72 score 2 dependents 50 scripts 585 downloads

rmorie - Multi-Domain Open Research and Inferential Estimation

Self-contained multi-domain scientific computing toolkit for observational inference and intervention analysis, hosting the Multilevel Reconciliation Methodology ('MRM') framework for Canadian correctional, police, and oversight data as its primary application. Every statistical algorithm is implemented in the package itself (R and 'C++') and cross-validated against its reference implementation; no other statistical package is called at runtime. Provides general-purpose causal estimators -- average treatment effects on the full, treated, and control populations ('ATE', 'ATT', 'ATC'), group, conditional, and local average treatment effects ('GATE', 'CATE', 'LATE'), augmented inverse-probability weighting ('AIPW'), and G-computation -- survey sampling methods (stratified, cluster, probability-proportional-to-size, bootstrap, calibration weights), propensity-score and doubly-robust estimators, and sensitivity analyses (E-value, 'Rosenbaum' bounds). Companion modules support signal processing and spectral analysis, hash and signature primitives, spatial statistics, statistical physics of crime ('Hawkes' self-exciting processes, reaction-diffusion, Levy flight, urban scaling), and classical-test-theory and item-response-theory measurement models, alongside ingestion utilities for officially published reports of the Ontario Special Investigations Unit ('SIU', <https://www.siu.on.ca/>, police oversight) and of federal Structured Intervention Units. Additional open-data connectors cover the 'CKAN' (Comprehensive Knowledge Archive Network, <https://ckan.org/>) portal family including Canada's federal portal (<https://open.canada.ca/>), Statistics Canada (<https://www.statcan.gc.ca/>), the municipal 'Socrata' portals of Chicago (<https://data.cityofchicago.org/>), New York City (<https://opendata.cityofnewyork.us/>), and Calgary (<https://data.calgary.ca/>), and the Toronto Police Service public-safety data portal (<https://data.tps.ca/>); none of these require an account or API (application programming interface) key.

Last updated

softwarestatisticalmethodsurvivalspatialtimeseriesbayesianmultiplecomparisonvisualizationacademic-researchcarceral-datacausal-inferencecriminologymrm-frameworkontario-siupolice-data-miningpolice-oversightsiusiu-parserspecial-investigations-unitcurllibsodiumopenblascppopenmp

5.86 score 63 scripts

rmoriedata - Integrated Datasets for the 'rmorie' Package

Integrated open data fixtures used by the 'rmorie' package for examples, vignettes, and tests. Contains snapshots of publicly available datasets from open-data portals built on the Comprehensive Knowledge Archive Network ('CKAN', <https://ckan.org/>), 'Socrata' (<https://dev.socrata.com/>), and 'Opendatasoft' (<https://www.huwise.com/>) (Chicago, New York City, Toronto, Vancouver, and others), Statistics Canada Canadian Centre for Justice and Community Safety Statistics ('CCJS') tables, a multi-agent-reviewed corpus of Ontario Special Investigations Unit ('SIU', <https://www.siu.on.ca/>) director's reports, and synthetic fixtures for unit tests. Also ships a small set of disclosure-control helpers for aggregate releases: differential privacy (an exact discrete Laplace sampler for counts and the analytic Gaussian mechanism for means, drawing on the operating system's random source, with privacy-budget accounting), k-anonymity and l-diversity checks, and cell suppression audited for exact recovery from the margins.

Last updated

softwareexperimentdataacademic-researchzlib

5.26 score 1 dependents 9 scripts 574 downloads