Skip to content
View Karant15's full-sized avatar

Block or report Karant15

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Karant15/README.md

Hi, I'm Karan Trivedi

Data Analyst | Healthcare · Pharma · Finance Analytics
MS Data Analytics — Webster University (Dec 2024) | Piscataway, NJ
Lean Six Sigma Black Belt | Python · SQL · R · Streamlit · Power BI

I build end-to-end analytics projects on real government and open datasets — from raw data pipelines through ML models and live interactive dashboards. All work is version-controlled with documented, reproducible code.


Current Role

Data Analyst — Phoenix Tech Solutions Inc. (Dec 2025 – Present, Remote)
Building a portfolio of sector-specific analytics solutions using public datasets to support new business proposals across healthcare, pharma, supply chain, and financial services.


Featured Projects

# Project What It Does Scale Stack Live
1 CMS Fraud, Waste & Harm Command Center Joined 5 government datasets to surface where pharma payments, opioid prescribing and patient harm intersect. 3 theses proven. 1.6M providers · 27GB Python · SQL · Streamlit Dashboard · Repo
2 Finance Fraud Detection XGBoost + SHAP fraud classifier on 590k real transactions. ROC-AUC 0.915. Live transaction risk scorer. 590k transactions XGBoost · SHAP · SMOTE · Streamlit Repo
3 Healthcare Workforce Analytics Mapped physician shortages across all 50 US states using 9.6M Medicare records. 9.6M records · 1.1M providers Python · Streamlit · CMS Dashboard
4 Healthcare Readmission ML Predicts 30-day hospital readmission risk. XGBoost + SHAP. Live risk predictor with gauge chart. 101k patient records XGBoost · SHAP · SMOTE · Streamlit Dashboard
5 Supply Chain KPI + DMAIC + SQL 57% late delivery rate found in 180k orders. 15 SQL queries. Full Six Sigma DMAIC analysis. 180k orders Python · SQL · SQLite · Streamlit Dashboard
6 SQL Business Analytics 25 SQL queries — basic aggregations through advanced window functions. Live SQL explorer in browser. 2 real datasets SQL · SQLite · Python · Streamlit Dashboard
7 Human Capital Analysis 97% turnover prediction accuracy. Decision Tree 0.97 AUC. 15,000+ employee records. 15k employees R · Logistic Regression · Decision Trees Repo
8 Bank Loan Default Risk 80.65% sensitivity. $1.165M cost reduction through threshold optimisation. 5,960 applicants R · Logistic Regression Repo
9 Consumer Segmentation K-Means + Random Forest segmentation for AXANTEUS market research agency. 600 profiles R · K-Means · Random Forest Repo

Project Highlights

🏥 CMS Fraud, Waste & Harm Command Center — LIVE

Sep 2026 · Biggest project

Joined 5 US government datasets (27GB total) — Open Payments, Medicare Part D, Medicare Part B, Hospital Compare, and the NPPES NPI Registry — using the NPI number as the primary key to build a master table of 1.6 million US providers.

Three theses proven with real data:

  • Opioid outliers prescribe at 113.5x the rate of low-scoring specialty peers
  • Brand drug outliers prescribe brand-name drugs 165% above specialty average
  • Top 6.4% of billing outliers account for 72.3% of all Medicare services billed

NJ Spotlight: 1,920 NJ providers · 213 high-risk · 477 opioid outliers

Dashboard: 6 interactive tabs — national risk map, 3 thesis views, NJ spotlight, provider lookup with risk gauge chart

🔗 Live Dashboard · GitHub Repo

Python Pandas NumPy SciPy SQLite SQL Streamlit Plotly


💳 Finance Fraud Detection Dashboard — LIVE

Aug 2026

End-to-end fraud detection pipeline on 590,540 real IEEE-CIS financial transactions.

  • XGBoost classifier with SMOTE class balancing — ROC-AUC 0.915
  • SHAP explainability layer for compliance team transparency
  • SQL intelligence layer — 6 queries including running total window functions
  • Live transaction risk scorer with fraud probability gauge

🔗 GitHub Repo

XGBoost SHAP SMOTE SQLite Streamlit Plotly


🏥 Healthcare Workforce Analytics — LIVE

Jan 2026

Analysed 9.6M real US Medicare records to surface physician staffing gaps.

  • 1.1M unique providers across 104 medical specialties
  • US choropleth map — state-level shortage classification
  • Wyoming, Vermont, Alaska identified as most critically underserved
  • DMAIC Six Sigma recruitment gap analysis

🔗 Live Dashboard


🏥 Healthcare Readmission ML Pipeline — LIVE

Mar 2026

Predicts 30-day hospital readmission risk from 101,745 real patient records.

  • 4 models trained and compared — XGBoost ROC-AUC 0.598
  • SHAP explainability surfaces top clinical risk drivers
  • SMOTE handles 11.2% class imbalance
  • Live risk predictor with gauge chart and clinical recommendations

🔗 Live Dashboard


📦 Supply Chain KPI + DMAIC + SQL — LIVE

Feb 2026

180,519 real orders — 57% delivered late across 23 global regions.

  • 15 SQL queries: aggregations, joins, subqueries, window functions
  • ABC inventory analysis — Class A drives 80% of revenue
  • Full DMAIC Six Sigma structured root cause analysis

🔗 Live Dashboard


🗄️ SQL Business Analytics — LIVE

Apr 2026

25 SQL queries across healthcare and supply chain real datasets.

  • Window functions: RANK, DENSE_RANK, ROW_NUMBER, LAG, LEAD, running totals, moving averages
  • CTEs for ABC classification and cross-domain joins
  • Live SQL Explorer — write and run any query in the browser

🔗 Live Dashboard


Tech Stack

Languages      Python · R · SQL
Visualization  Plotly · Streamlit · Power BI · Tableau · Seaborn
ML             XGBoost · SHAP · SMOTE · Scikit-learn · Random Forest
               Logistic Regression · Decision Trees · K-Means Clustering
Database       SQLite · PostgreSQL · MySQL · SQL Server · DAX
               Relational Modelling · Dimensional Modelling
Process        Lean Six Sigma Black Belt · DMAIC · SIPOC · RCA · FMEA
Domain         Healthcare · Pharma · Finance · Supply Chain · HR Analytics
Tools          Git · GitHub · Jupyter · VS Code · Power BI Desktop

Certifications & Education

  • MS Data Analytics — Webster University (Dec 2024) · GPA 3.31
  • Lean Six Sigma Black Belt — Benchmark Six Sigma (2021)
  • Lean Six Sigma Green Belt — Benchmark Six Sigma (2021)
  • Google Data Analytics — Coursera
  • Microsoft PL-300 Power BI — Microsoft Learn

Prior Experience

Senior Accounts Manager — ID Medical LLP (Nov 2016 – Jul 2021)
Managed 30+ NHS hospital accounts across the UK · Forecasting accuracy +25% · 15% YoY revenue growth · Supported a CMS-focused healthcare sourcing initiative — directly relevant to the CMS analytics work above

Senior Recruitment Consultant — QX KPO Services (Aug 2014 – Oct 2016)
453 shifts filled in one month · £25,000 revenue · Led a 4-member team

International Peer Mentor / Writing Coach — Webster University (Jan 2023 – Dec 2024)
CRLA Level 2 Certified · 94% positive outcome rate


Contact

📫 krntrivedi@gmail.com
🔗 LinkedIn
🔗 CMS Command Center · Healthcare Analytics · Fraud Detection · Supply Chain

Pinned Loading

  1. CMS-Command-Center CMS-Command-Center Public

    Joining 5 CMS government datasets (27GB) to surface where pharma payments, opioid prescribing, and patient harm intersect - 1.6M providers, 3 proven theses, live Streamlit dashboard

    Python

  2. Healthcare-ML-Pipeline Healthcare-ML-Pipeline Public

    End-to-end ML pipeline predicting 30-day hospital readmission - XGBoost, SHAP explainability, FastAPI, Streamlit

  3. Healthcare-Workforce-Analytics Healthcare-Workforce-Analytics Public

    Analyzing 9.6M US Medicare records to identify physician staffing gaps and recruitment priorities across 50 states

  4. HR-Attrition-ML-Pipeline HR-Attrition-ML-Pipeline Public

    Predicting employee attrition using XGBoost and SHAP explainability - 1470 real IBM HR records - SQL analysis - live Streamlit dashboard

    HTML

  5. Human-Capital-Analysis Human-Capital-Analysis Public

    Predicting employee turnover using Logistic Regression, Decision Tree, k-NN & SVM on 14,999 employees. Decision Tree achieved 97% accuracy & 0.97 AUC. Built in R. Dataset: 10 attributes.

    R