Analyze transaction data to identify fraud patterns and risk indicators using SQL and Python.
This project simulates real-world fraud analysis scenarios, focusing on transaction monitoring, anomaly detection and risk evaluation.
- SQL (PostgreSQL)
- Python (Pandas)
- Jupyter Notebook
- Identification of suspicious transactions based on value and frequency
- Creation of fraud detection rules
- Calculation of anomaly rate
- Exploratory Data Analysis (EDA)
- High-value transactions concentrated in short time intervals indicate risk patterns
- Repeated transactions from same user increase fraud probability
- Behavioral patterns are critical for fraud detection
- Clone repository
- Open notebooks
- Run analysis
- Applied SQL and Python in real fraud scenarios
- Developed analytical thinking for risk identification
- Built structured data analysis pipeline