wtf is credit-bureau-analysis?

This project is a portfolio piece that walks through a full data analytics workflow using a made up credit bureau dataset for a Nigerian financial institution. The goal is to show every step an analyst would take, starting with generating synthetic data, cleaning it, analyzing it with SQL, and finally presenting the findings in an executive Power BI dashboard. The dataset contains ten thousand customer credit records spanning ten Nigerian states and three years, from 2022 through 2024, and it includes several planted data quality problems for practice, such as duplicate rows, missing values, and inconsistent spelling in fields like gender and employment status. The project documents how each of these issues was found and fixed, for example replacing missing numeric values with the median rather than the average, since a few extreme outliers would otherwise distort the results. Twelve SQL queries, ranging from simple counts to more advanced window functions and scoring logic, answer business questions about the overall default rate, which states and loan types carry the most risk, and which customers are close to defaulting but have not yet done so. The analysis found an overall default rate near thirty three percent, with Rivers State standing out at over sixty percent, nearly double the national average, and payday loans showing the highest share of customers falling behind on payments. The final output is a single page Power BI dashboard called A Portfolio Under Pressure, combining summary metric cards, charts breaking down default rates by state and risk category, and a table listing customers who are at risk of default soon. The project folder includes the raw and cleaned datasets, the SQL query file, the Power BI file itself, and supporting documentation describing the data dictionary and the cleaning steps taken.