Overview

Banks process thousands of loan applications every month — but turning raw loan, customer, and repayment data into actionable insights requires an integrated analytics pipeline.

This project demonstrates an end-to-end analytics workflow that connects

MS SQL Server (data layer) → Python (ETL + Machine Learning) → Power BI (reporting layer).

It uses a synthetic but realistic dataset representing loan applications, repayments, and customer details, and builds a pipeline that cleans data, predicts loan approval likelihood, and visualizes portfolio performance.

The analysis focuses on:

• Loan portfolio health (approval probability, missed payments)
• Customer insights (income, employment, loan purpose)
• Feature engineering & model prediction
• Operational dashboarding using Power BI

🧩 Problem Statement

The bank’s analytics team needed an automated system to answer:

• Which new applications are likely to be approved or high risk?
• How can we quantify missed payments and loan-to-income ratios efficiently?
• Can we combine SQL + Python + Power BI for a reproducible, refreshable pipeline?
• How do we monitor model performance and update predictions dynamically?

To address this, I designed a Loan Analytics Pipeline that unifies all data and analytics steps in one reproducible workflow.

🧰 Tools & Technologies