{"id":23688878,"url":"https://github.com/anidipta/risk-sense","last_synced_at":"2026-01-30T14:20:21.594Z","repository":{"id":289824597,"uuid":"822092721","full_name":"Anidipta/Risk-Sense","owner":"Anidipta","description":"Risk Sense : Next-Gen Fraud Predictor","archived":false,"fork":false,"pushed_at":"2025-01-01T15:49:32.000Z","size":28516,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-09-02T16:47:29.211Z","etag":null,"topics":["convolutional-neural-networks","machine-learning","machine-learning-algorithms","numpy","python","streamlit"],"latest_commit_sha":null,"homepage":"https://risksense.streamlit.app","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Anidipta.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","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,"zenodo":null}},"created_at":"2024-06-30T09:42:53.000Z","updated_at":"2025-04-16T13:20:54.000Z","dependencies_parsed_at":"2025-04-25T09:06:29.495Z","dependency_job_id":null,"html_url":"https://github.com/Anidipta/Risk-Sense","commit_stats":null,"previous_names":["anidipta/risk-sense"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/Anidipta/Risk-Sense","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Anidipta%2FRisk-Sense","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Anidipta%2FRisk-Sense/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Anidipta%2FRisk-Sense/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Anidipta%2FRisk-Sense/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Anidipta","download_url":"https://codeload.github.com/Anidipta/Risk-Sense/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Anidipta%2FRisk-Sense/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28914052,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-30T12:13:43.263Z","status":"ssl_error","status_checked_at":"2026-01-30T12:13:22.389Z","response_time":66,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":["convolutional-neural-networks","machine-learning","machine-learning-algorithms","numpy","python","streamlit"],"created_at":"2024-12-30T00:19:49.964Z","updated_at":"2026-01-30T14:20:21.584Z","avatar_url":"https://github.com/Anidipta.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 📊 **Risk Sense : Next-Gen Fraud Predictor**\n\n## **Dataset Overview**\nThis dataset is specifically designed for developing and evaluating machine learning models focused on **fraud detection** in financial transactions. It contains **6.3 million** rows of simulated transactional data, offering a comprehensive foundation for building and testing models to detect fraudulent activities.\n\n## **Columns in the Dataset:**\n\n1. **⏳ `step`**: Represents a unit of time where 1 step equals **1 hour**.\n2. **💳 `type`**: The **type** of transaction, including the following categories:\n   - `CASH_IN`\n   - `CASH_OUT`\n   - `DEBIT`\n   - `PAYMENT`\n   - `TRANSFER`\n3. **💰 `amount`**: The **monetary value** of the transaction.\n4. **🏦 `oldbalanceOrg`**: The initial **balance of the origin account** before the transaction.\n5. **🏦 `newbalanceOrig`**: The updated **balance of the origin account** after the transaction.\n6. **🏦 `oldbalanceDest`**: The initial **balance of the destination account** before the transaction.\n7. **🏦 `newbalanceDest`**: The updated **balance of the destination account** after the transaction.\n8. **🔄 `changebalanceOrg`**: The **change in balance** for the origin account following the transaction.\n9. **🚨 `isFraud`**: A **binary indicator** (0 or 1) where:\n   - `1`: Fraudulent transaction\n   - `0`: Non-fraudulent transaction\n\n---\n\n## **Dataset Summary:**\n\n- **📈 Total Rows**: 6,300,000\n- **📊 Total Columns**: 9\n\n### **Key Features:**\n- **🔢 Transaction Types**: A variety of transaction types, making the dataset suitable for modeling different fraud scenarios.\n- **💳 Balance Changes**: Tracks balance changes before and after transactions for both origin and destination accounts.\n- **🚨 Fraud Indicator**: Essential for supervised learning, helping models identify fraudulent behavior.\n\n---\n\n## Flow-Case\n\n![Flowchart](https://github.com/Anidipta/Risk-Sense/blob/main/Images%20Model/Flow.png)\n\n---\n\n\n## **Potential Use Cases:**\n\n1. **🔍 Supervised Learning for Fraud Detection**: \n   - Train models to predict the likelihood of fraudulent transactions.\n   - Develop predictive models using **binary classification** techniques such as decision trees, XGBoost, or neural networks.\n\n2. **📉 Pattern Analysis**: \n   - Analyze transaction patterns that indicate potential fraud.\n   - Discover features such as **unusual amounts** or rapid **balance changes** that could suggest fraudulent activities.\n\n3. **🛠️ Feature Engineering**: \n   - Create new features to improve the performance of models, such as:\n     - Transaction frequency over time.\n     - Account balance changes relative to previous transactions.\n     - Time-based behavior patterns for accounts.\n\n---\n\n## **How to Run:**\n\nAccess our **final fraud detection model** [here](https://drive.google.com/file/d/1P2HRWjud5vZ3E5PRUhqvywu9UHo8xttO/view?usp=sharing).\n\n### **Live Demo**: \nNo download is needed! **Try the fraud detection model live** at [Risk Sense](https://risksense.streamlit.app/).\n\n---\n\n## **Data Source \u0026 Collection Method:**\n\nThe dataset is a **simulated representation** of real-world financial transactions. Each row represents a transaction with attributes designed to mimic actual banking behavior. The simulation includes both **fraudulent** and **non-fraudulent transactions**, providing a diverse environment for model training and evaluation.\n\n---\n\n## **Challenges and Considerations:**\n\n- **⚖️ Class Imbalance**: \n  Fraudulent transactions are significantly fewer than non-fraudulent ones, which could result in **model bias** towards predicting non-fraudulent transactions. This can be mitigated by:\n  - **Resampling techniques** such as **SMOTE** or **ADASYN**.\n  - Using **ensemble models** that can better handle imbalance, like **Random Forest** or **XGBoost**.\n\n- **🔒 Data Privacy**: \n  While this dataset is simulated, it mimics the structure of real-world transactional data, which can still be useful for creating privacy-preserving algorithms for **real-time fraud detection** systems.\n\n---\n\n## **Conclusion:**\n\nThis **Fraud Detection Dataset** is a comprehensive resource for developing robust fraud detection models. With a variety of features capturing transaction behavior and the fraud indicator, it provides ample opportunities for **pattern recognition**, **anomaly detection**, and **predictive modeling**. This dataset is ideal for both **academic research** and **industry applications** aiming to enhance financial security through automated fraud detection.\n\n---\n\n## **Dependencies:**\n\nEnsure you have the following Python libraries installed:\n\n```bash\ncatboost\nstreamlit\npandas\nnumpy\njoblib\nstreamlit_lottie\nscikit-learn\n```\n\n---\n\n## **Demo Video**: \nFor an introduction to the fraud detection system, watch the demo video [here](https://youtu.be/qHkBchgEdTg?si=mCmb0Dm8TBo88reV).\n\n---\n\n## **Author Information**\n\n| **Name** | **Year** | **Position** |\n|:---:|:---:|:---:|\n| **Anidipta Pal** | 1st | Data Engineer, Data Analyst, ML Engineer |\n| **Sagnik Basak** | 1st | Full Stack Developer |\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fanidipta%2Frisk-sense","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fanidipta%2Frisk-sense","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fanidipta%2Frisk-sense/lists"}