Skip to content

Security: radyslav-datascience/substitution_coefficient_validation

Security

SECURITY.md

Security & Data Privacy

Overview

This repository contains research code and aggregate results for pharmaceutical substitution coefficient analysis. This document describes what data is present, what has been excluded, and why.

What Is Published

Content Status Notes
Analysis scripts (study_*/scripts/) ✅ Published No sensitive data
Configuration modules (config/, core_parameters/) ✅ Published Thresholds, paths, algorithm parameters
Sequential algorithm (determining_markets_algorithm/) ✅ Published Production-ready code, no data
Aggregate result tables (correlations, deviations, distributions) ✅ Published Statistics only — no drug identifiers
Visualization charts (PNG) ✅ Published Statistical distributions, no drug names
Written conclusions and methodology docs ✅ Published Aggregate summaries only
Excel reports with aggregate statistics ✅ Published No per-drug name-value pairs
Study documentation (docs/) ✅ Published Methodology, parameters, context

What Is Excluded

Content Reason
Raw coefficient files (data/sp4_coefficients.csv, data/sequential/) Contain drug names (DRUGS_NAME) paired with per-market substitution values
Drug classification data (data/classification/) Drug names paired with classification labels and confidence metrics
Per-drug comparison tables (cross_study_comparison_*.csv) Drug names paired with cross-study numerical metrics
Methods research data (methods_research/*/data/) Drug names with GMM/clustering coefficient results
Pipeline output data (pipeline/data/output/) Drug names with final computed coefficients per segment
Study 03b per-drug results (per_drug_comparison.csv, data/output/) Drug names with cross-variant stability metrics

Pharmacy & Market Identifiers

All pharmacy locations and local markets are referenced exclusively by internal numeric IDs (e.g., CLIENT_ID = 108139). These IDs cannot be used to identify individual pharmacies without access to proprietary registry data.

Drug Information

Drug names (commercial and INN) appear in excluded data files only — always paired with quantitative research results. Published content (charts, aggregate tables, conclusions) does not expose individual drug names alongside their performance metrics.

No Personal Data

This research contains no personally identifiable information (PII). All data represents aggregated pharmacy sales transactions at the drug-SKU level. No patient, prescriber, or consumer data is present at any stage.

Reporting Security Issues

If you identify a file in this repository that you believe should not be publicly accessible, please contact:

There aren't any published security advisories