{"id":14982341,"url":"https://github.com/soumyadeepbose/fraud-detector","last_synced_at":"2026-01-07T07:16:05.512Z","repository":{"id":250592049,"uuid":"831829830","full_name":"soumyadeepbose/fraud-detector","owner":"soumyadeepbose","description":"Project FraudCatch leverages AI to predict and prevent financial fraud in real-time. It uses Apache Kafka for data streaming, Apache Spark for distributed processing, and differentially private machine learning models for security. Our Streamlit application provides businesses with real-time fraud analysis, notifications, and preventive measures.","archived":false,"fork":false,"pushed_at":"2024-08-01T13:20:29.000Z","size":294,"stargazers_count":1,"open_issues_count":0,"forks_count":1,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-02-14T10:41:51.179Z","etag":null,"topics":["apache-kafka","apache-spark","aws","aws-s3","finance","financial-data","firebase-auth","python","streamlit","streamlit-dashboard","streamlit-webapp"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/soumyadeepbose.png","metadata":{"files":{"readme":"readme.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.txt","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-07-21T18:17:15.000Z","updated_at":"2024-08-01T13:20:32.000Z","dependencies_parsed_at":"2024-09-29T06:15:24.828Z","dependency_job_id":null,"html_url":"https://github.com/soumyadeepbose/fraud-detector","commit_stats":{"total_commits":23,"total_committers":2,"mean_commits":11.5,"dds":0.04347826086956519,"last_synced_commit":"dfc6e45ccef5539a5f592f72bfcb935e1a3f9c4f"},"previous_names":["soumyadeepbose/fraud-detector"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soumyadeepbose%2Ffraud-detector","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soumyadeepbose%2Ffraud-detector/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soumyadeepbose%2Ffraud-detector/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/soumyadeepbose%2Ffraud-detector/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/soumyadeepbose","download_url":"https://codeload.github.com/soumyadeepbose/fraud-detector/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245940259,"owners_count":20697357,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["apache-kafka","apache-spark","aws","aws-s3","finance","financial-data","firebase-auth","python","streamlit","streamlit-dashboard","streamlit-webapp"],"created_at":"2024-09-24T14:05:14.023Z","updated_at":"2026-01-07T07:16:05.482Z","avatar_url":"https://github.com/soumyadeepbose.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Project Fraudcatch `alpha` 😊\n\n## Info Edge Ventures AI Hackathon 2024 - Team Scamslayers💀💀✨\n\n### Team Details\n- **Team Name:** Scamslayers💀💀✨\n- **Team Members:** [Sampriti Mitra](https://github.com/Sampriti03) and [Soumyadeep Bose](https://github.com/soumyadeepbose)\n\n### Problem Statement\nUse of AI to predict and prevent financial frauds.\n\n### Project Overview\nWe have created a production-level fraud detection and prevention system for financial transactions for the InfoEdge Ventures AI Hackathon 2024. Our system is built using a data pipeline, distributed model inference, differential privacy, and continual incremental learning.\n\n### Solution Components\n1. **Data Ingestion:**\n   - Real-time data streaming using Kafka from the company’s API.\n   \n2. **Model Inference:**\n   - Distributed model inference using Apache Spark to predict fraudulent transactions.\n   - Ensemble of several models to increase accuracy.\n   \n3. **Data Storage:**\n   - Storing the data in an S3 bucket upon successful inference using differential privacy.\n   \n4. **Analysis and Dashboard:**\n   - Analyzing stored data and displaying results on a Streamlit dashboard.\n   \n5. **User Interface:**\n   - A Streamlit website for businesses to sign up, connect their API, access fraud data analysis, and receive fraud prevention measures.\n   - Real-time notifications for fraudulent transactions.\n   \n6. **Security and Privacy:**\n   - Using differential privacy techniques for data protection and equitable analysis.\n   - Reporting false positives to continually train the models.\n\n### Key Features\n- **Real-Time Anomaly Detection:** Immediate detection and response to fraudulent activities using Kafka.\n- **Distributed Data Processing:** Using Spark for distributed processing and inference.\n- **Streamlit Application:** Users can upload data, connect their API, and access an analysis dashboard.\n- **Real-Time Notifications:** Email alerts to security teams upon detection of fraud.\n- **Scalability and Flexibility:** The system is designed to handle large volumes of data and adapt to various fraud detection needs.\n- **Differentially Private Models:** Ensuring user data remains confidential while maintaining high accuracy in fraud detection.\n\n### Technologies Used\n- **Data Streaming:** Confluent Kafka, Confluent Zookeeper\n- **Data Processing:** Bitnami Spark Master, Bitnami Spark Worker\n- **Storage:** AWS S3\n- **Machine Learning:** Tensorflow, Pytorch, Scikit-learn, Diffprivlib\n- **Web Framework:** Streamlit\n- **Programming Languages:** Python\n- **Other Tools:** Mlflow, Docker, Pandas, Matplotlib\n\n### Performance Metrics\n- **Ensembled Models Accuracy:** 94% accuracy on the test set.\n- **Average Spark Processing Delay:** Less than 1 second.\n- **Kafka Latency:** 50-80 milliseconds.\n- **Data Retrieval Time:** 100 milliseconds per record.\n\n### Future Enhancements\n- Implementing homomorphic encryption or secure multi-party computation.\n- Integrating blockchain technology for logging transactions and model updates.\n- Extending fraud detection capabilities beyond credit card fraud.\n- Adding user activity tracking with Google Firebase.\n- Deploying the system through AWS ECR for scalability.\n- Developing explainability features for real-time fraud detection insights.\n\n### GitHub \u0026 Demo Video\n- **GitHub Repository:** [fraud-detector](https://github.com/soumyadeepbose/fraud-detector)\n- **Demo Video:** [Watch Demo](https://drive.google.com/file/d/1izJ7p9-Au0fZBDVPAhrIiRB8sdAkQ1no/view?usp=sharing)\n\n### Project Architecture\n![Architecture Diagram](website/images/Project_FraudCatch_Alpha.svg)\n\n---\n\nThis project was developed for the Info Edge Ventures AI Hackathon 2024 by Team Scamslayers.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoumyadeepbose%2Ffraud-detector","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsoumyadeepbose%2Ffraud-detector","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsoumyadeepbose%2Ffraud-detector/lists"}