Open to opportunities

Abantika Basu

@abantikabasu

I’m a finance-focused data analyst using Python, SQL, and econometrics to turn messy data into predictive, risk-aware insights.

India

Message

What I'm looking for

I’m looking for a Data Analyst role where I can apply Python/SQL and econometrics to build predictive models, explain drivers with SHAP, and support finance and risk-focused decisions with clear, business-ready insights.

I’m a finance-specialized economics graduate building strong, data-driven models across fraud detection, churn prediction, and econometric analysis. Through my Data Analyst Intern roles, I cleaned and aggregated real-time production and market data with Python, then built a fraud detection system using classification and anomaly detection—achieving 85–90% precision and reducing false negatives using SMOTE and a Random Forest approach.

I also develop end-to-end predictive analytics solutions: a churn model delivered 85–90% accuracy with ROC AUC > 0.85 and used SHAP to explain the top churn drivers, while a housing price pipeline achieved R² 0.85–0.90 and reduced error by 15–25%. Beyond ML, my projects analyze policy and social data using UDISE+ and econometric methods (OLS/Logit, ARDL), including identifying teacher shortages across 1M+ schools and finding long-run relationships in Nifty returns.

Experience

Work history, roles, and key accomplishments

HR Intern

Bharat Economic Forum

Jul 2025 - Oct 2025 (3 months)

Supported HR operations by onboarding new interns and coordinating with internal teams. Helped host investor meetings and assisted communication between investors, teams, and management.

Meeting Coordination Team Support Administrative Coordination Time Management

Data Analyst Intern

Infitech Learn

Jun 2025 - Jul 2025 (1 month)

Developed end-to-end machine learning projects focused on predictive analytics and model optimization. Built a customer churn model with Random Forest achieving 85–90% accuracy and ROC AUC > 0.85, and used SHAP to identify key churn drivers; also improved housing price prediction with 3 models (R² 0.85–0.90) and reduced error by 15–25%.

Machine Learning Predictive Analytics Random Forest ROC AUC SHAP Optimization Churn Modeling Data Analysis

Data Analyst Intern

KnackTech

Nov 2024 - Jan 2025 (2 months)

Performed real-time production and market data analysis, including data cleaning and aggregation. Built a fraud detection system using classification and anomaly detection, achieving 85–90% precision and reducing false negatives; improved minority-class detection by 35–40% using SMOTE and Random Forest.