A Consideration Point of Fraud Detection in Bank Loans - Project Aug 2018
-
Updated
Feb 18, 2019 - Jupyter Notebook
A Consideration Point of Fraud Detection in Bank Loans - Project Aug 2018
Determing the eligibility for granting home loan. ML classification models are used, in order to predict if loans are apporoved or not, based on customers's data.
🏦 Project IBRD Statement of Loans (Python, SQL, Excel, Power BI, Tableau)
An interactive Power BI dashboard that analyzes bank loan data to provide insights into approval trends, default risks, and customer profiles. Designed to assist financial institutions in making data-driven lending decisions.
A house loan dashboard created using Microsoft Power BI
Repository of my general analytics project about bank personal loan modelling.
Exploratory data analysis on bank loan data from 90,000 customers using Python code, SQL queries, and Tableau visualizations.
AI-powered DPR and project report generator for PMEGP, NLM and other government schemes, with subsidy calculations, financial projections, Excel-based formulas, and professional bank-ready reports.
Interactive Bank Loan Analysis Dashboard built in Tableau. Monitors $459M in funded loans, analyzing credit risk through grade-wise distribution, verification status, and regional repayment trends.
Predict loan approvals instantly with explainable AI, complete end-to-end ML pipeline, demo app, and FastAPI deployment.
To associate your repository with the bank-loans topic, visit your repo's landing page and select "manage topics."