# FOSS_Hackathon 2026

🏦 FinRisk — AI-Powered Borrower Risk Analyzer (Multi-Device, Backend-Driven)

A full-stack fintech web application for intelligent loan risk assessment, designed for FOSS Hackathon 2026 with a focus on open-source contribution, real-world impact, and robust architecture.

---

## 📌 About FinRisk

FinRisk is a real-time, AI-powered loan risk analysis platform that connects borrowers, bankers, and admins in a seamless digital loan workflow. It integrates RBI/CIBIL-standard scoring, alternative data risk assessment, government scheme discovery, and future financial planning.

Now with backend-based data storage, users can login from multiple devices and see consistent, real-time financial and loan data.

---

## Key Features

### 👤 Borrower

* Secure Registration & Login – Multi-device, session-based authentication.

* Financial Input Dashboard – Income, expenses, EMIs; data stored securely in backend.

* Loan Eligibility Checker – Enhanced 5-factor scoring model:

* Standard Factors: FOIR, Credit Score, Savings Rate, LTI, Age

* Alternative Data: Social footprint, transaction history patterns, digital footprint → feeds risk score

* Apply for Loan – Submit requests with tenure, type, and age.

* Real-Time Status Updates – Notifications when banker picks, approves, or rejects.

* Loan & Expense Tracker – Upload bills, categorize expenses, view breakdown charts.

* Financial Advisory – AI-driven recommendations based on risk score and savings trends.

* Government Schemes – Access 12 official schemes (APY, PMJDY, PMJJBY, MUDRA) with working links.

* Future Planning & Skill Development – Set goals; 15 curated courses from NPTEL, Google, SEBI, AWS, CFA.

### 🏦 Banker

* Banker Registration – Branch, IFSC, Banker ID.

* Loan Request Dashboard – Pick unassigned requests.

* Decision Workflow – Approve / Reject / Manual Review (with reasons).

* Freeze Action Buttons – Prevent accidental duplicate decisions.

* Performance Analytics – Approval %, rejection %, total processed.

### ⚙ Admin

* Overview Dashboard – All banks, borrowers, and bankers.

* Detailed Analytics – Individual and overall charts, performance metrics.

* Live System Stats – Landing page shows active borrowers, pending loans, approvals.

---

## 🧠 Risk Scoring Model

Hybrid scoring using standard and alternative data sources:

| Factor | Weight | Source |

| --------------------------------------- | ------ | ------------------------------ |

| CIBIL / Credit Score | 35 pts | TransUnion CIBIL |

| FOIR (Fixed Obligation to Income Ratio) | 25 pts | RBI Circulars / NBFC norms |

| Net Savings Rate | 20 pts | Income - Expenses (User Input) |

| LTI (Loan to Annual Income) | 12 pts | Calculated from financial data |

| Age Factor | 8 pts | RBI/NHB Lending Norms |

Hard Knockouts:

* CIBIL < 550 → capped at 30 pts

* FOIR > 65% → capped at 25 pts

Alternative Data Risk Integration:

* Uses digital footprint, utility payments, and micro-transactions

* Supplements traditional metrics for borrowers without formal credit history

Final Risk Score = Weighted Standard Score + Alternative Data Score

Enables *more inclusive lending** for underbanked users

---

## 🛠 Tech Stack & Architecture

Backend: Python 3.11, Flask

Database: SQLite3 (multi-device support)

Frontend: HTML5, CSS3, JavaScript, Chart.js

Fonts: Google Fonts (Syne, DM Mono)

Authentication: Session-based, JWT-compatible for multi-device login

Data Flow: Backend stores all user, loan, and analytics data → ensures consistent multi-device experience

Notes:

All interactions pass through backend → ensures *secure, synchronized data**.

* Risk Scoring Engine runs in backend → outputs risk scores visible to borrower, banker, admin.

* Notifications handled via backend → real-time status updates.

* Charts & analytics generated via frontend pulling backend JSON endpoints.