AstraFin - Your financial present and future

Imagine having a financially savy version of yourself with an ML-powered decision-making strategy that learns your spending habits, boosts savings, and helps you achieve your goals with high accuracy.

Comment

Inspiration

Managing personal finances can be overwhelming, especially when tracking multiple accounts and spending categories. We wanted to create an intelligent, user-friendly platform that helps people understand their financial behavior and make better decisions. AstraFin was born from the idea of combining financial data with machine learning to give users insight, foresight, and control all in one place. We are also heavily focused on accesibility by allowing users to talk to their smart financial companion which is built from you past transactions.

What it does

AstraFin is an all-in-one banking application that uses AI and ML to simplify financial management.

  • Provides users with a clear overview of their current financial situation.
  • Categorizes expenses automatically using intelligent transaction analysis.
  • Generates personalized insights about spending patterns.
  • Predicts future account balances using ML models trained on user history.
  • Offers data-driven recommendations to improve savings and budgeting habits.
  • Allows users to set their savings goals and gives guidance to reach those goals in a clear, concise manner.

How we built it

AstraFin’s backend is built with Python and NodeJS, handling API integration, machine learning models, and secure database management.

  • Python: Implements ML pipelines and predictive analytics while also hosting out API endpoints.
  • NodeJS: Manages server-side logic and user requests.
  • PostgreSQL: Stores transaction histories, user profiles, and financial data securely.
  • React + ViteJS: Powers a responsive, fast, and intuitive frontend interface.
    The system communicates through RESTful APIs, ensuring modularity and scalability across all components.

Challenges we ran into

  • Integrating multiple services while maintaining consistent data flow between backend and frontend.
  • Designing a robust ML pipeline that adapts to user-specific financial patterns.
  • Ensuring accurate balance predictions while minimizing overfitting on limited datasets.
  • Managing real-time updates and state synchronization between the database and user interface.
  • Balancing technical complexity with usability to keep the app accessible to non-technical users.
  • Providing accessibility at the apps core. Everyone should be able to access, use and test our application regardless of their capabilities.

Accomplishments that we're proud of

  • Successfully developed an ML model that predicts account balances with high accuracy.
  • Built a seamless integration between financial data, analytics, and an engaging UI.
  • Created a scalable architecture that can support future AI-driven features.

What we learned

  • The importance of clean data preprocessing in financial prediction tasks.
  • How to structure a scalable full-stack application that blends AI, data analysis, and modern web technologies.
  • Best practices for securely handling sensitive financial data and API communication.

What's next for AstraFin

  • Expanding predictive capabilities to include investment forecasting.
  • Integrating external financial APIs for real bank synchronization.
  • Launching a mobile app version for cross-platform accessibility.
Share this project:

Updates