How thrinesh13/medicare-claims-processing-analytics works
This repository is an analytical prototype for evaluating Medicare Part B claim straight-through processing, not an operational claims adjudication service. Its workflow generates synthetic claim lifecycle data, derives time-available features in PostgreSQL, scores claims and compares manual, checks-only, and checks-plus-score policies by accuracy and decision cost. The README also documents calibration, validation, sensitivity, and reporting analyses. The graph follows the pipeline and sampled Python/SQL responsibilities; data files referenced by source but absent from the supplied tree are not represented as repository nodes.
Main entry points: evaluate.py Scoring and policies · database.py Database access · 02_decision_economics.ipynb Decision economics analysis · run_pipeline.py Pipeline entry point
Components 20
Claim Data
- CMS source retrieval
scripts/fetch_sources.py - Claim lifecycle generator
src/generate.py - Pipeline entry point
scripts/run_pipeline.py
Feature Engineering
- Database access
src/database.py - Claims schema
sql/01_schema.sql - As-of features
sql/02_features.sql
Policy Evaluation
- Scoring and policies
src/evaluate.py - Fitted model
results/model.joblib
Economics & Analysis
- Decision reporting
sql/03_reporting.sql - Policy summaries
results/policy_summary.csv - Cost sensitivity
results/sensitivity.csv - Calibration results
results/calibration.csv - Monthly results
results/monthly.csv - Segment results
results/segments.csv - Claim decisions
results/fact_decisions.csv - Data and SQL analysis
notebooks/01_data_and_sql.ipynb - Decision economics analysis
notebooks/02_decision_economics.ipynb
Other components
- Analyst
- CMS data
- PostgreSQL
Connections 28
- CMS source retrieval → CMS data: requests data
- Pipeline entry point → Claim lifecycle generator: calls
- Pipeline entry point → Database access: loads and features
- Pipeline entry point → Scoring and policies: runs evaluation
- Claim lifecycle generator → CMS source retrieval: uses benchmarks
- Database access → Claims schema: executes
- Database access → PostgreSQL: loads claims
- Database access → As-of features: executes features
- As-of features → PostgreSQL: queries history
- Database access → Scoring and policies: returns features
- Scoring and policies → Fitted model: saves
- Scoring and policies → Policy summaries: writes
- Scoring and policies → Cost sensitivity: writes
- Scoring and policies → Calibration results: writes
- Scoring and policies → Monthly results: writes
- Scoring and policies → Segment results: writes
- Scoring and policies → Claim decisions: writes
- Database access → Decision reporting: executes reporting
- Database access → Claim decisions: loads decisions
- Decision reporting → PostgreSQL: creates reporting table
- Analyst → Data and SQL analysis: explores
- Analyst → Decision economics analysis: evaluates results
- Data and SQL analysis → PostgreSQL: queries features
- Decision economics analysis → Policy summaries: reads
- Decision economics analysis → Cost sensitivity: reads
- Decision economics analysis → Calibration results: reads
- Decision economics analysis → Monthly results: reads
- Decision economics analysis → Segment results: reads