Data Analyst | SQL · Python · Relational Databases & Data Warehousing MS Data Analytics | Webster University (Dec 2024) | Piscataway, NJ
I build end-to-end analytics projects — from SQL extraction through dashboards and ML models — using public/open datasets across healthcare, supply chain, HR, and finance. All work here is version-controlled with documented, reproducible code.
Data Analyst — Phoenix Tech Solutions Inc. (Dec 2025 – Present, Remote)
Assigned to build a portfolio of sector-specific analytics solutions using public and open-source datasets, to support new business proposals across healthcare, supply chain, human capital, and financial services. The projects below were built as part of this work.
| # | Project | What It Does | Stack | Date | Live |
|---|---|---|---|---|---|
| 1 | Healthcare Workforce Analytics | Analyzed 9.6M Medicare records to map physician shortages across 50 states | Python · Streamlit · CMS Data | Jan 2026 | Open |
| 2 | Supply Chain KPI + DMAIC + SQL | Found 57% late delivery rate in 180K orders. 15 SQL queries. DMAIC framework. | Python · SQL · SQLite · Streamlit | Feb 2026 | Open |
| 3 | Healthcare Readmission ML | Predicts 30-day hospital readmission. XGBoost + SHAP. 101K real patient records. | XGBoost · SHAP · SMOTE · Streamlit | Mar 2026 | Open |
| 4 | SQL Business Analytics | 25 SQL queries — basic to advanced window functions. Live query explorer in browser. | SQL · SQLite · Python · Streamlit | Apr 2026 | Open |
| 5 | Human Capital Analysis | 15,000+ employee records. 97% prediction accuracy for turnover (Decision Tree, 0.97 AUC). | R · Logistic Regression · Decision Trees | May 2026 | Repo |
| 6 | Bank Loan Default Risk Model | 5,960 applicants. 80.65% sensitivity. $1.165M cost reduction through threshold optimization. | R · Logistic Regression | Jun 2026 | Repo |
| 7 | Consumer Segmentation Analytics | 600 customer profiles segmented for AXANTEUS market research agency. | R · K-Means · Random Forest | Jul 2026 | Repo |
Jan 2026
Analyzed 9.6M real US Medicare records to identify physician staffing gaps across all 50 states.
- Processed 1.1M unique providers across 104 medical specialties
- Built 5-tab Streamlit dashboard with US choropleth maps
- Applied Lean Six Sigma DMAIC framework to recruitment gap analysis
- Wyoming, Vermont, and Alaska identified as most critically underserved
- Live: https://karan-healthcare-analytics.streamlit.app
- Stack: Python · Pandas · Plotly · Streamlit · CMS Medicare Data
Feb 2026
180,519 real orders analyzed — 57% delivered late across 23 global regions.
- 42.7% on-time delivery rate — Central Africa worst at 60.7%
- 15 SQL queries covering late rates, revenue, customer segments
- ABC inventory segmentation — Class A drives 80% of revenue
- Full DMAIC Six Sigma structured analysis
- Live: https://karan-supply-chain.streamlit.app
- Stack: Python · SQL · SQLite · Plotly · Streamlit · DMAIC
Mar 2026
End-to-end ML pipeline predicting 30-day hospital readmission risk.
- 101,745 real patient records from 130 US hospitals
- 4 models trained and compared — XGBoost won with ROC-AUC 0.598
- SMOTE to handle 11.2% class imbalance
- SHAP explainability layer surfaces clinical decision rationale
- Live risk predictor with gauge chart and clinical recommendations
- Live: https://karan-healthcare-ml.streamlit.app
- Stack: Python · XGBoost · SHAP · SMOTE · Streamlit
Apr 2026
25 advanced SQL queries across healthcare and supply chain real datasets.
- 6 basic, 9 intermediate, 9 advanced queries
- Window functions: RANK, DENSE_RANK, ROW_NUMBER, NTILE, LAG, LEAD, running totals, moving averages
- CTEs for ABC inventory classification and cross-domain state comparison
- Live SQL Explorer — write and run any SQL against a real database in browser
- Live: https://karan-sql-analytics.streamlit.app
- Stack: SQL · SQLite · Python · Plotly · Streamlit
May 2026
15,000+ employee records. Repo
- 97% prediction accuracy for turnover (Decision Tree, 0.97 AUC)
- Job satisfaction and compensation identified as primary retention drivers
- Stack: R · Logistic Regression · Decision Trees
Jun 2026
5,960 applicants. Repo
- Improved model sensitivity to 80.65%
- $1.165M cost reduction through threshold optimization
- Stack: R · Logistic Regression · Decision Trees
Jul 2026
600 customer profiles segmented for AXANTEUS market research agency. Repo
- K-Means clustering, logistic regression, random forest
- Identified high-value, brand-loyal customer segments
- Stack: R · K-Means · Random Forest
Languages: Python · R · SQL (GROUP BY through window functions and CTEs)
Visualization: Plotly · Streamlit · Power BI · Tableau · Seaborn
ML: Scikit-learn · XGBoost · SHAP · Random Forest · Logistic Regression
Decision Trees · Clustering · Predictive Modeling
Imbalance: SMOTE (imbalanced-learn)
Database: SQL · SQLite · PostgreSQL · MySQL · DAX · Relational & Dimensional Modeling
Process: Lean Six Sigma Black Belt · DMAIC · SIPOC · RCA · FMEA
Domain: Healthcare · Supply Chain · HR Analytics · Finance · Recruitment
Version Control: Git/GitHub
- Lean Six Sigma Black Belt — Benchmark Six Sigma (2021)
- Lean Six Sigma Green Belt — Benchmark Six Sigma (2021)
- MS Data Analytics — Webster University (Dec 2024), GPA 3.31
- Google Data Analytics — Coursera (in progress)
- Microsoft PL-300 Power BI — Microsoft Learn (in progress)
Senior Accounts Manager / Client Manager — ID Medical LLP (Nov 2016 – Jul 2021) 30+ NHS hospital accounts · Forecasting accuracy +25% · 15% YoY revenue growth · Supported a federal CMS-focused sourcing initiative
Senior Recruitment Consultant — QX KPO Services (Aug 2014 – Oct 2016) 453 shifts booked 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 across mentored cohort
📫 krntrivedi@gmail.com | LinkedIn 🔗 Healthcare Dashboard · Supply Chain Dashboard · ML Pipeline · SQL Analytics