Inspiration
The inspiration for SmartStore is focused on a key challenge in today's energy landscape: the increasing changes of energy markets. The project idea came about after observing how large facilities struggle to optimize energy costs and decision-making. Problem Statement: Large facilities face significant challenges in managing their electricity costs and environmental impact. Traditional energy consumption patterns rely heavily on grid power during peak demand periods when electricity prices are highest, leading to inflated operating costs and increased carbon emissions. While battery storage systems offer a potential solution, facility managers lack the sophisticated tools and insights needed to optimize their energy storage investments. They struggle to make real-time decisions about when to charge batteries, when to use stored energy, and how to balance cost savings with system longevity.
How we built it
Our development process began with a workflow and breakdown structure divided into three main components: frontend UI development using HTML/CSS for user input and display, backend implementation featuring ML algorithms and training models, and data analytics integration with EIA API. The frontend team focused on creating an user-friendly interface while the backend team developed the ML infrastructure, with the data analytics team establishing connections to energy price data sources and APIs.
Challenges we ran into
We faced significant challenges in finding and integrating a suitable energy API that could provide reliable real-time price data for our model. The integration of our frontend dashboard with backend services was another challenge. Additionally, we overcame deployment issues on Render and made efforts in creating an interface that could effectively communicate complex energy data to non-technical users.
What we learned
Our team gained invaluable experience in integrating complex systems, particularly in combining ML algorithms with real-time data processing and creating an user-centric frontend interface for energy trading. We also learned the importance of effective and optimized team collaboration and agile development, allowing us to overcome technical challenges and deliver a functional MVP within the hackathon's limited time window.
Accomplishments that we're proud of
We successfully developed an energy trading platform that combines real-time data with ML-powered predictions, achieving a remarkable 91.6% accuracy in our initial model testing. Our team effectively integrated multiple components: the dashboard to backend ML algorithms.
What's next for SmartStore
We plan to expand SmartStore's capabilities by scaling to multiple geographic regions, integrating government APIs, and enhancing our ML models with more diverse datasets for improved prediction accuracy. We plan to optimize for sub-millisecond trading response times and implementing advanced visualization tools to provide users with visualization of market insights.
Log in or sign up for Devpost to join the conversation.