Inspiration
During the 2021 power crisis in Texas, millions of Americans watched as their neighbors and loved ones suffered through days-long blackouts in freezing temperatures. Without access to heat, hundreds lost their lives during the failure of the larger power grid. Overreliance on the power grid puts the population’s safety at the mercy of its integrity, raising serious concerns, as seen in the tragedy in the Lone Star State. Motivated by this feeling of helplessness, our team was determined to create a tool that provides our communities with greater energy independence. We hoped to provide a second layer of protection during grid failures while also reducing the burden on the power grid.
What It Does
Throughout this weekend, our group focused on creating an application that is able to optimize solar panel models and locations based on an address, budget, desired energy output, and other criteria.
What We Learned
Throughout the process, we developed new technical skills that helped bring our project to life. This constant learning was especially present when working with existing APIs and making changes to our model to better fit the purpose of our application. Our team also solidified our GitHub fundamentals, preparing us for future competitions.
How We Built It
To bring our idea to life, we implemented a system of refinement, using AI coding tools to create a basic model, which we then built upon to fix errors and further tailor it to our goal. In particular, we used Anthropic's Claude and OpenAI's Codex for guidance and assistance with higher-level coding techniques.
We connected our application to the Google Solar API, which allowed us to analyze rooftops and locate suitable areas for solar panel installation. Additionally, we adjusted the model to fit our parameters, adding different solar panel types to better match each solution to the user’s needs. We then added other supporting tools, including an interactive map system along with an interface to compare different types of panels.
Challenges We Ran Into
While creating our application, we had to make hard decisions regarding the balance between accuracy and speed. When creating more intricate algorithms, the model would take
Log in or sign up for Devpost to join the conversation.