HomeRec

Logo

Inspiration

The inspiration behind HomeRec was to create a equal opportunity platform where users of any background can utilze our technologies in researching viable investment opportunities

What it does

With HomeRec, users can make informed decisions on the homes that they can invest in. Specifically, we utilize a cosine similarity test to identify the most profitable and affordable option for home buyers based on their query results. We also attempted to implement a dashboard where users can see their pinned homes that they wish to buy.

How we built it

We built it by separating into group of frontend and backend developers to accelerate development.

Challenges we ran into

There were no readily available APIs or databases for housing, so we decided to make our own database, which is composed of 50 houses in the Blacksburg, Christiansburg, and Roanoke areas and supplemental information (area, bedrooms, bathrooms, price, etc.), which was used for our search feature. We also took the 10 year price history of each house using Zillow's online Zestimates and used it to train and test our machine learning model to predict real estate pricing.

Accomplishments that we're proud of

Implemented working OAuth, usable card design for each home, search feature, made two custom datasets (one for search feature other for time series model), Utilized ARIMA model to predict future home prices

What we learned

We learned a lot of valuable information involving the full development process as well as learning what worked for us to be efficient and make a product. We learned and explored new technologies, such as mongoDB, propelAuth, and forming an ARIMA model.

What's next for HomeRec

With time, a housing API could be used to further populate our search function and further train our ML model. The search function could also benefit with a filtering aspect, allowing users to narrow down houses based on budget and other preferences.