← Back to Projects

FutureSpend — See Tomorrow, Save Today, Share Success

AI-powered personal finance dashboard that predicts upcoming spending from calendar events and transaction data. Connects your calendar to spending predictions: analyzes upcoming events (meals, outings, transport), predicts likely spend by category, and surfaces insights, savings challenges, and an AI coach. Built with Next.js, TypeScript, Tailwind, and a FastAPI backend deployed on Render.

FutureSpend financial dashboard with spending forecasts, category insights, and budgeting metrics
Financial dashboard with spending forecasts, category insights, and budgeting metrics. Calendar-driven pipeline surfaces upcoming expenses and recommended actions.

Problem & Context

Calendar-driven financial awareness is often missing: people see their week’s spend only after it happens. FutureSpend addresses this by turning your calendar into a spending forecast — built in 24 hours at a hackathon sponsored by SFU CSSS and RBC. The system analyzes upcoming events (dining, social, transport, entertainment), predicts likely spend by category using a rules-based pipeline, and surfaces insights and recommended actions (e.g. trim one event to stay under budget). Target users: individuals and teams who want a lightweight, calendar-first view of upcoming expenses.

What It Does

Tech Stack

Next.js 14 React 18 TypeScript Tailwind CSS Recharts D3 (Sankey) Python 3.11 FastAPI Uvicorn Pydantic Google Gemini Render GitHub Pages

Architecture / How It Works

Two-tier setup: Frontend (Next.js 14 with output: 'export' for static GitHub Pages) and Backend (FastAPI + uvicorn on Render). The frontend talks to the backend via NEXT_PUBLIC_API_URL. Calendar events (mock or Google) are parsed into features, run through a prediction pipeline for category-based spend, then feed the dashboard, insights, challenges, and optional Gemini-powered AI coach. Banking and leaderboard endpoints support the demo experience.

Key Takeaways

FutureSpend demonstrates calendar-first personal finance: predicting spend before it happens and making small trade-offs. The modular stack — Next.js static export, FastAPI backend, optional Gemini integration — is built for hackathon speed and can be extended with real calendar and bank APIs. Working with event parsing, category-based prediction, and an AI coach in a single 24-hour build provided hands-on experience in full-stack deployment (Render + GitHub Pages) and user-focused financial UX.

← Back to Projects