-
-
Dashboard
-
Dashboard Part 2
-
Landing Page
-
Example Redacted Bill
Track Chosen
Smart Grid
Inspiration & Problem Statement & Ideation
Users do not generally prioritize thier residential carbon footprint as they usually do not know thier carbon footprint. Currently, utility bills generally show the amount of electricity used, and the monetary cost of that electricity. If the bill is too high, then people generally take steps to try to reduce their electricity bill. Following a similar theme, we created OhmSweetOhm, which will show users the carbon price of their electricity usage, in hopes that users would take steps to reduce it if it is too high.
What it does, Solution Proposed and Intended Impact
OhmSweetOhm takes a .pdf of the user's utility bill and uses Gemini to parse important data such as zip code and kWhs of electricity used. Don't worry, all PII is safely redacted!
We then utilize NREL's end-use load profiles for US homes, and based on the user's home information, we select an appropriate energy profile. This data is paired with corresponding grid mix data for the region to determine whether the user's house is more or less carbon intensive that the reference NREL home data.
This data is also passed to Gemini to get personalized tips on reducing the user's residential carbon footprint!
We hope that by showing users this data, and tracking it, users are able to make data-driven decisions about managing their residential carbon footprint.
How we built it (Development Process)
We used a React frontend and a partly Gemini-powered Flask backend. We extract the grid mix data from the US Energy Information Administration, and the Residential End Use Load Profile data from NREL. This data is aggregated and used to built a representative hourly load profiles for users based on their single utility bill. We can then track, compare, and suggest improvements for users to reduce their residential carbon footprint.
Challenges we ran into
We had a lot of trouble getting the frontend and backend to communicate effectively, and had to spend many hours debugging various related issues that arose when trying to put the application together. Also, due to the scale of data available, it was tempting to make the application available to the entire USA. However, we had to maintain our scope, and do the few things we set out to do, right.
Accomplishments that we're proud of
We got the frontend and backend working seamlessly with each other, sending the complex data needed for the visualizations on our dashboard. We dialed in the login and security with Auth0, and set up effective prompts for our Gemini work. All in all, the application works as intended.
What we learned
We learnt a lot about residential load profiles, and about the terabytes of residential and commercial load profile data hosted by NREL on AWS's S3. We also got a good bit of data science and full-stack development experience.
What's next for OhmSweetOhm
We'd like to expand OhmSweetOhm to cover the entire United States. We also want to build out our features to include long term carbon footprint tracking, as well as more detailed analytics on the user's residential carbon usage. We would also like to allow users to upload thier own smart meter data, to get even more personalized load profiles and improvement feedback.
Log in or sign up for Devpost to join the conversation.