This project analyzes credit card transaction data to identify fraud patterns using Python, Pandas, and Tableau.
A Python-based ETL (Extract, Transform, Load) pipeline processes the dataset before building an interactive dashboard to explore fraud trends and anomalies.
🔗 Live Dashboard:
https://public.tableau.com/app/profile/praneel.reddy.kanduri/viz/fraud_detection_dashboard_twbx/CreditCardFraudDetectionAnalysis
file:///Users/kandurivineelreddy/Desktop/Credit%20Card%20Fraud%20Detection%20/creditcard_dashboard.html
The dataset contains 284,807 credit card transactions, including 492 fraudulent transactions (~0.17%), making it a highly imbalanced dataset.
This project focuses on analyzing transaction behavior to identify patterns associated with fraudulent activity.
Key objectives:
- Process raw transaction data using a Python ETL pipeline
- Perform exploratory data analysis
- Build a Tableau dashboard to visualize fraud patterns
- Identify anomalies in transaction amounts and time patterns
- Python
- Pandas
- Tableau
- Git & GitHub
The ETL pipeline consists of three stages:
Loads the dataset using Pandas.
Performs:
- Data inspection
- Fraud vs normal transaction analysis
- Dataset structure exploration
Exports processed data for visualization in Tableau.
The Tableau dashboard includes three main analyses:
Fraud vs Normal Transactions
Visualizes the class imbalance between legitimate and fraudulent transactions.
Fraud Transactions Over Time
Shows temporal spikes in fraud activity.
Transaction Amount Distribution
Highlights differences in transaction amounts between fraud and normal transactions.
- Fraud transactions represent ~0.17% of the dataset
- Fraud tends to occur in short bursts over time
- Fraudulent transactions show greater variability in transaction amounts
Praneel Reddy Kanduri
Aspiring Data Analyst | Python | SQL | Tableau