Detecting Fraud Patterns in Credit Card Transactions: A Data Science Approach

Project Overview

This project analyzes credit card transactions to detect fraud patterns using Exploratory Data Analysis (EDA). The goal is to identify key trends, anomalies, and insights in fraudulent vs. non-fraudulent transactions.

Dataset:

• Source: Kaggle - Credit Card Fraud Detection
• Transactions: 555,719
• Fraud Cases: 2,145 (0.39%)

Key Objectives:

• Understand the distribution of fraud vs. non-fraud transactions
• Analyze time-based patterns in fraudulent activity
• Identify high-risk transaction types
• Create interactive visualizations with Tableau

Key Findings:

• Fraud transactions often involve higher amounts.
• Fraud is more frequent at night.
• Fraud transaction values increase during the day.

Key Steps:

1. Data Preprocessing:

   • Cleaned and processed the raw dataset by removing unnecessary columns and handling missing values.
   • Creating new features based on existing variables

2. Exploratory Data Analysis (EDA):

   • Explored the distribution of fraud vs. non-fraud transactions.
   • Analyzed transaction amounts and frequencies to identify patterns.
   • Investigated time-based trends in fraudulent transactions.
   • Univariate and bivariate analyses were performed to identify potential fraud indicators.

3. Tableau Visualization:

   • Built interactive dashboards in Tableau to showcase the findings visually, focusing on transaction patterns, fraud detection over time, and high-risk transaction types.
   • Added filters for customer history to allow deeper insights:
     • Age: Analyze trends in fraud across different age groups.
     • Sex: Explore the distribution of fraud incidents by gender.
     • Location: Filter by geographic regions to identify regional fraud patterns.
     • These filters enable users to explore how age, sex, and location impact fraudulent behavior.

Project Structure

.
├── data                              # Folder containing the raw and processed data
│   ├── fraudTest.csv                 # The dataset used for fraud detection analysis
│   └── README.md                     # Data description and usage guide
├── images                            # Folder for storing any visual assets
├── reports                           # Folder to store project reports and analyses
│   └── Data Analysis & Insights.pdf  # Report on data analysis and insights gained from the project
├── scripts                           # Folder containing Python scripts used for preprocessing, analysis and visualization
│    ├── fonts                        # Folder for storing the custom font files used in the visualizations
│    │    └── Montserrat-Regular.ttf  # Custom font Montserrat-Regular
│    ├── __init__.py                  # Marks the 'scripts' directory as a Python package
│    ├── data_check.py                # Script for checking the dataset for inconsistencies or issues
│    ├── data_cleaning.py             # Script for cleaning and preprocessing the dataset (e.g., handling missing data)
│    └── eda_functions.py             # Script containing functions used in exploratory data analysis (EDA) and visualization
├── tableau                           # Folder to store Tableau workbooks and dashboards
│    └── Fraud_detection_insights.twb # Tableau workbook containing fraud detection insights and interactive dashboards
├── .gitattributes                    # Defines Git attributes like line endings, diff settings, etc.
│    └── data/*.csv                   # Use Git LFS for CSV files in the data folder
├── .gitignore                        # Specifies files/folders for Git to ignore (e.g., temporary files, credentials)
│    └── node_modules/                # Ignore node_modules directory
│    └── *.log                        # Ignore log files
│    └── __pycache__/                 # Ignore Python bytecode cache
│    └── *.pyc                        # Ignore Python bytecode files
├── LICENSE                           # License file containing the full license terms
├── README.md                         # The readme file with project overview, objectives, and instructions
├── main.ipynb                        # Main script to run the project
└── requirements.txt                  # List of Python dependencies required for the project (e.g., pandas, matplotlib, etc.)

How to Run the Project

1. Install Dependencies

Before starting, make sure you have all the necessary dependencies installed. You can install them by running the following command in your terminal:

pip install -r requirements.txt 

2. Run the Jupyter Notebook:

Open main.ipynb in Visual Code and execute the cells sequentially:

main.ipynb 

Technologies Used

• Python (Pandas, Matplotlib, Seaborn)
• Visual Studio Code
• Tableau (for interactive visualizations) - Dashboard
• Trello Board (for project management) - Data Analytics Project

Next Steps

• The next phase of the project will involve using machine learning algorithms to build predictive models for detecting fraud.
• After selecting the best-performing model, the next step is to deploy it for real-time or batch predictions on new transactions.

Author

Developed by Annie Meneses Gonzalez.

Feel free to connect via LinkedIn or check out more projects on GitHub and Tableau

This is an open-source project and contributions are welcome. Please ensure that you follow the MIT license guidelines while contributing to this project.

License

This project is licensed under the MIT License.