Inspiration
We're broke college kids and want to know how much things cost before buying it.
What it does
Our website displays a Price Predictor based on data from the Costar dataset using machine learning.
How we built it
We used pandas library to parse through dataset and host the information using the streamlit library, which abstracts the front-end of our app. In reading in data through a .csv file, we attempting to optimize performance by manually allocating the minimum required amount of bytes for each dtype column. We also used pytorch to make and train a neural network that takes the dataset and some parameters and outputs the predicted price.
Challenges we ran into
First time collaborators and had to learn a lot about the group project, common libraries, and git functions. (ie. lots of merge conflicts)
Accomplishments that we're proud of
We actually got a working website! For a first time competing, we're happy.
What we learned
Come to the event with a plan and organization so that most of the preparation is done ahead of time so the 36 hours used can be spent on actually coding and making progress. Some never used python before.
What's next for DC Realestate Price Predictor
We want to try to implement a map such as from Google maps to locate specific ones. We can use a GPU instead of a laptop cpu to better train the neural network.
Log in or sign up for Devpost to join the conversation.