Northwind Pulse

Inspiration

Utility complaints often reveal problems only after customers have already been affected. We wanted to help teams spot billing risks earlier, resolve complaints more effectively, and use what they learn to guide future investments.

What we built

Northwind Pulse brings those steps into one workflow. It surfaces risky bills, helps teams track and resolve complaints, and summarizes outcomes. Its Investment Planner uses Gemini to recommend evidence-backed actions, while deterministic backend calculations estimate costs, savings, ROI, and payback.

How we built it

We built the interface with Next.js, React, and TypeScript, with a FastAPI backend that analyzes Northwind complaint and utility datasets. Gemini proposes and explains strategy options; the backend handles the financial calculations. We label planning assumptions clearly, and identify synthetic account-level data as demo data.

Challenges and what we learned

A key challenge was combining AI recommendations with financial outputs users can trust. We kept those responsibilities separate: Gemini recommends, while transparent calculations use the selected assumptions and available data. We also learned how important it is to show data provenance and limitations alongside projected results.

Links

The link to the github repository containing our project is included in this submission, along with a google drive link to our 1 page value case.

Built With

• fastapi
• google-gemini-api
• next.js
• pandas
• python
• react
• tailwind-css
• typescript