A recent trend has emerged from the tech industry: platforms that simplify and allow the public to participate in investing. Apps such as Robinhood and Acorn have taken the investment community by storm, racking up millions of users in months. We can see that investing is no longer just for people who have thousands or millions of dollars to spend. However, these apps are geared toward equity market trading, ignoring one of the largest sectors of investing - real estate. Our application, Bedrock, is a simple to use application that fills that gap and allows the public to dip their toes into the world of real estate investing.

What it does

Property owners are always looking for funding for development projects and plans. Through Bedrock, developers can put their properties up for funding and our algorithms help pair up these properties with individuals of the general public who want to invest in real estate. Bedrock also uses machine learning algorithms to help users determine the potential of different development projects. Our backend analytics uses numerous factors such as lot size, square feet, and property type to produce an evaluation of the property. Users can compare the developer's asking price with our evaluation to judge if it's worth investing into. Another useful tool is the list of keywords we generate from the notes of buildings using natural language processing. Users can use these keywords to obtain a better qualitative description of the property.

How we built it

Bedrock is split into three primary sections: frontend, backend, and machine learning. The machine learning aspect covers the scikit learn neural network and Google Cloud's natural language processing API. We ported results into javascript so our backend can calculate the evaluations and generate the keywords. Backend took data from our dataset along with the calculated evaluation and stored it within a MongoDB database. We also used the Google Maps api to allow investors to see where the properties are. Frontend focused on retrieving information from the database in order to display graphics, property information and other factors.

Our Accomplishments

Our web application was a platform that allowed both property owners and investors to buy and sell equity in real estate. Not only that, but our algorithms also predicted housing prices in order to provide investors with an evaluation of the property so that property owners won't label it fraudulently. Our NLP algorithm also provided interesting insight into keyword patterns and could be used by both property owners and investors to see how they are trending.

What we learned

MongoDB does not work well with Google Cloud. Porting neural networks into javascript without flask is difficult.

What's next for Bedrock

A wider range of portfolio analytics and graphics. Taking inflation into account when processing our evaluation of properties. Working functionality of selling real estate equity.

Share this project: