- 📍 Based in Jaipur, India
- 🎯 Seeking Data Analyst roles
- 🔍 I turn messy data into actionable business decisions
- 🛠️ Building end-to-end analytics projects with Python, SQL, Power BI and Tableau
| Category | Tools |
|---|---|
| Languages | Python, SQL |
| Libraries | Pandas, Matplotlib, Seaborn, Scikit-learn |
| BI Tools | Power BI, Tableau |
| Other | Excel, Git, GitHub |
Python · Pandas · SQL · Scikit-learn · K-Means · Cosine Similarity
- Segmented customers into clusters — Cluster 3 (15% of base) drives 40% of total revenue
- Built lookalike model using cosine similarity for high-value customer targeting
- Identified Cluster 5 as highest churn risk — longest gap since last purchase, lowest AOV
- SQL queries for revenue by cluster, Q4 seasonality analysis, at-risk customer flagging
- Key insight: 65% of high-value customers are from Southeast Asia — recommended as next fulfillment expansion market
Python · Pandas · SQL · SQLite · Tableau
- Analyzed attrition across department, job role, age group, overtime and tenure
- Sales Representatives have the highest attrition at 40%
- Overtime employees leave at 30% vs 10% for non-overtime — 3x higher risk
- Highest attrition in years 0–2 — onboarding identified as biggest retention failure
- 📊 View Tableau Dashboard
Power BI · DAX · Excel
- Tracked $244K revenue across 250 orders and 714 units sold
- Electronics top category (140K) · Refrigerator top product (75K) · Miami top city (32K)
- PayPal most used payment method (24%) · Sales peaked Feb 2025 then declined sharply
- Key insight: 65% of orders are pending or cancelled — critical fulfillment gap identified
Python · Pandas · Tableau · GitHub Pages
- Analyzed 500 rows across 20+ countries, 10 genres, Free vs Premium users
- Classical leads global streams · R&B has longest avg stream duration
- Free and Premium users skip at nearly identical rates (20%) — skipping is content-driven, not ads
- South Korea, Sweden, South Africa lead total hours streamed
- 🌐 Live Dashboard · 📊 Tableau Public
Python · SARIMA · Holt-Winters · Streamlit
- Implemented SARIMA and Holt-Winters models with automated parameter selection
- ADF stationarity checks, seasonal decomposition, missing value imputation
- Achieved MAE: 2.34 · RMSE: 3.15 · MAPE: 4.5%
- Deployed interactive Streamlit interface with date range selection and forecast horizon control
"Data is only useful when it drives a decision."
