NIESCHAL GHOSH

Money Mule Detection Model

• Developed a machine learning model to detect mule accounts at an early stage.
• Achieved a high Gini score of 0.95, significantly enhancing the bank’s ability to prevent fraudulent activities.

Mutual Fund Cross-Sell Model

• Built a logistic regression model to identify potential customers for mutual fund cross-selling.
• Achieved a Gini score of 0.71, supporting targeted marketing and improving campaign efficiency.

ATM Off-Us Hits Forecasting Model

• Developed a time series forecasting model using SARIMA to predict off-us ATM transaction volumes.
• Enabled more accurate resource planning and optimization of ATM network operations.

Branch Location Optimization

• Conducted pincode-level analysis to identify optimal locations for new branch openings.
• Leveraged Google Earth and web scraping to gather and analyze geospatial and demographic data, supporting data-driven decision making.

Customer Segmentation for High-Spend Categories

• Built models to identify customer segments with high spending in supermarket and travel & tourism categories.
• Enabled targeted marketing strategies by uncovering behavioral and demographic patterns.