WeSave

Inspiration

As international students, we often found ourselves struggling with managing our finances. We could never pinpoint the root cause of our expenses, and credit card usage only added to the complexity. WeSave was born from the desire to gain financial clarity while also contributing to the community.

What it does

WeSave is a financial analysis tool that helps users track their spending habits, identify major expense categories, and predict future expenditures. Additionally, it provides AI-driven insights to help users make smarter financial decisions. A small percentage of the user’s financial gains is automatically redirected to support low-income individuals, fostering a cycle of community support.

How we built it

WeSave leverages several technologies to provide a seamless experience. The frontend is built using Next.js, Tailwind CSS, and shadcn. The backend is powered by Flask, utilizing various APIs, including SaltEdge for transaction data, Gemini AI for financial insights, and Firebase for authentication and database storage.

Challenges we ran into

We faced multiple challenges throughout development.

1. Machine Learning Model Optimization
   Training and fine-tuning the expense prediction model took hours of trial and error, from selecting the right algorithms to hyperparameter tuning.

2. AI Integration with Gemini
   Structuring financial transaction data in a way that Gemini AI could understand without misinterpreting prompts as casual conversation was an ongoing struggle.

3. Secure Transaction Data Retrieval
   Initially, we considered web scraping bank data, but it posed security and reliability issues. We also explored parsing email receipts, but not all banks provide structured transaction data. Ultimately, we opted for SaltEdge, a secure and industry-compliant API for accessing financial data.

Accomplishments that we're proud of

• Successfully integrating SaltEdge to securely retrieve and analyze real-time financial data.
• Developing an AI-driven expense prediction system that offers actionable insights.
• Creating a seamless donation process to automatically contribute to low-income users.
• Designing a user-friendly and intuitive interface for financial tracking.

What we learned

• The importance of financial literacy and how small habits can lead to significant savings.
• Advanced AI prompt engineering to refine responses from Gemini.
• Securely handling sensitive financial data while maintaining user trust.
• The complexities of machine learning model optimization for real-world financial applications.

What's next for WeSave

• To provide cheaper alternatives to users expensive expenditures.
• Have a chrome extension that can look for impulse purchases and warn the user before they make the purchase
• Create SMS notifications for users who may potentially be late on their credit card statements

Built With

• email
• firebase
• flask
• gemini
• nextjs
• python
• saltedge
• tailwind
• typescript
• vercel