Real Estate price Predictor based on Computer Vision

Inspiration

We are a team full of students of Computer Science, very keen about the Machine and Deep Learning that we are living now.

What it does

Based on a Decision Tree, we predict the approximate value of a house, using 4 photos of different parts (frontal exterior, bedroom, bathroom and kitchen) along with another attributes like the square meters or zipCode.

How we built it

The DecisionTree model is built with Scikit Learn library. And, with a backend of Flask (Python framework) and a HTML+CSS+JS frontend we built a web app to introduce the necessary inputs to predict about the price.

Challenges we ran into

We experienced a lot of challenges, like finding a appropiate Dataset and extract the features with de relstb.ai API based on Computer Vision, and training the AI Model. Another challenge was build the entire app, also remarkable the server backend built.

Accomplishments that we're proud of

We are really proud about the style that our web has, with a elegant and sophisticant touch that looks so good. Besides, all what happens behind like the features extraction of the dataset and the model training.

What we learned

We learned how to built an entire web app (frontend and backend) in different languages that we barely had touched. Also, searching along the entire net trying to find a appropiate dataset of real estate, that cost us the initial hours of the project.