ZotNest

Allows the User to input the details of their ideal apartment to get their ideal match ranked score
ZotNest shows the world the best options ranked based on their preferences with the closest to 100 and green being the best
Able to select specific apartments to get more information
Shows a full screen view of their floorplan
Price slider allows you to reduce the options shown

Inspiration

Finding student housing is a complex task of looking through corporate websites and real estate platforms such as Zillow. While these websites seem useful, they don't realize what students care about, such as social life and proximity. Through Zotnest, we ranked choices based on students' needs, making the stress of finding a new home simplified.

What it does

Our web application gives a curated score for each floor plan and community based on the user queries. Additionally, it provides background on why the llm chose this location and characteristics of the property in an interactive map. This streamlines the housing hunt process.

How we built it

Our web application consists of 4 main components: frontend, backend, database, and an LLM. First, the frontend gets the user query and sends a POST request to the backend. The backend receives it and the query is sent to the LLM. Our product works like an RAG system, where the LLM uses information in the database to generate results (consisting of characteristics and their reasoning) and a score. The backend then receives the results from the llm and brings them to the React frontend through FastAPI. Finally, the frontend displays the results to the user which consists of all the properties sorted by score.

Challenges we ran into

Challenges that we ran into were bringing the backend information into the correct frontend router and getting an accurate score. These issues were later resolved.

Accomplishments that we're proud of

We are proud of the final product because at end everything integrated smoothly, creating a good user experience. Additionally, we are proud of our idea because it is something that has not been done before and would be incredibly useful for students.

What we learned

From this project, we learned about the Retrieval Augmented Generation System and its benefits in providing curated results to users. We also learned how to create API endpoints to create a functioning front-end and back-end integration.