Explore Projects ↓
Built a Streamlit app using XGBoost to predict loan defaults on 1.3M+ records. Achieved 67% recall, 0.43 F1 score. Key risk drivers: FICO <660, DTI >35%, 60-month terms, and loan >₹12L.
Deployed Streamlit app using logistic regression (84% accuracy, 0.81 F1) to predict telecom churn. Found 55% churn in monthly plans, 3× risk without tech support, and 50%+ for tenure <6 months.
Analyzed 6.3M+ PaySim transactions using SQL to detect fraud. Uncovered ₹12B+ in flagged flows, 0.13% fraud rate, and peak activity during 6–11 PM. Built queries to score high-risk users and transactions.
Cleaned a global layoffs dataset using MySQL Workbench — handled NULLs, duplicates, and standardized formats for analytics readiness.
Conducted exploratory data analysis on layoffs — derived trends, aggregations, and country-wise patterns using advanced SQL queries.
Interactive Excel dashboard with slicers, pivot tables and charts analyzing bike sales across regions, categories and timelines.
A Power BI dashboard visualizing global data professionals’ survey responses — showcasing salary, roles, tool usage and geography insights.
Python + BeautifulSoup scraper to collect and clean revenue data of top U.S. firms from Wikipedia for competitive analysis.
Analyzed customer reviews using Tableau — visualizing sentiment trends, word clouds and key themes across feedback data.
I'm a final‑year B.Tech student passionate about solving real‑world problems through data. With hands‑on experience in SQL, Python, and Streamlit, I specialize in building end‑to‑end analytics and machine learning apps that deliver real business value. Not just a fresher — but a deployment‑ready analyst, with live projects and production‑grade skills.
"Turning raw data into strategic decisions."