About the project

Inspiration

Low-income households in the UK are disproportionately affected by energy price volatility, yet they lack access to the tools and insights that traders use to anticipate these changes. I wanted to bridge this gap by transforming complex market signals into simple, actionable advice that helps vulnerable users make better energy decisions in real time.

What it does

SaveMyEnergy is a decision engine that combines electricity market signals with predictive modelling to generate personalised, real-time advice such as “use later” or “save £X”. It translates complex price forecasts into clear, accessible recommendations, prioritising simplicity and usability for low-income households.

How I built it

The frontend was built using React (Vite) to create a clean, accessible interface. The backend uses FastAPI to handle prediction logic and decision-making. A lightweight forecasting model estimates short-term energy prices, which are then classified into pressure levels (low/medium/high). Based on this, the system selects optimal appliance usage times and calculates potential savings. I integrated OpenRouter for AI-generated advice and used APScheduler to simulate real-time updates and notifications.

Challenges I ran into

One of the main challenges was balancing accuracy with simplicity - ensuring the system was technically sound while still being understandable for non-technical users. I also faced issues with deployment (handling monorepo structure and environment configuration) and ensuring consistent behaviour between local and production environments. Additionally, estimating realistic energy prices without relying on hardcoded values required careful consideration.

What I learned

I learned how to design systems that translate complex data into meaningful real-world impact, especially for vulnerable users. I also gained experience in full-stack development, API integration, and deploying applications using modern tools like Vercel. Most importantly, I learned the importance of accessibility and clarity when building technology intended for real-world decision-making.

What’s next

I aim to integrate real-time UK energy market APIs for more accurate pricing, improve personalisation using household-specific data, and expand accessibility features (e.g., voice guidance, multilingual support). Ultimately, I want to turn this into a scalable platform that empowers households to actively manage and reduce their energy costs.

Built With

• apscheduler-(background-tasks)
• fastapi-(python)
• git-&-github
• html/css
• javascript
• openrouter-api-(llms)
• react-(vite)
• rest-apis
• smtp-(email-notifications)
• vercel