SunPredict

Inspiration and What it does

It shows you whether you should use your battery power, sell it, or save it for another day, helping save money, balance the load on the grid, all while using weather data and other cues to do so.

How we built it

We used the languages below to create it, and also we trained an ML model to match the data.

Challenges we ran into

NOAA broke 5 hours before the deadline, leaving us with essentially no base. Hard to find out good ideas for an MVP.

What we learned

How to create and tune a machine learning model as well as how to utilize APIs on demand and combine them to create usable data.

What's next for SunPredict

Waiting for a NOAA/Weather.gov fix, enhancing models, building more features, like a rotatable solar panel where you can see how much light it absorbs.

Built With

• cdo
• flask
• javascript
• ml
• noaa
• node.js
• svelte
• tensorflow
• typescript