Real Estate Predictor

Our User Interface.

Inspiration

The ability to predict future housing growth given current events will enable real estate companies to target those specific areas to maximize gains.

What it does

Analyze current event headlines with machine learning algorithms, correlate that data to housing growth, and assemble a model that can predict future growth given a set of data.

How we built it

We used a text sentiment analyzer to rank the headlines on a scale from 1 to 5 using a library called TextBlob. We then used Keras, a library on top of Tensorflow to create a machine learning neural network with 1 hidden layer. Our input feature was the sentiment of the text and our output feature was the expected change in Case/Shiller Home Price Index in DC. We used pandas to process the data and used Keras' built-in evaluation metrics to evaluate the model. We used HTML and CSS for the webpage, with the intention of using Flask to run our Python code on the back end.

Challenges we ran into

Our biggest challenge was our lack of data. It was very difficult to find news articles corresponding to each month from 20+ years ago. We also had trouble integrating everything together with Flask because we ran out of time.

Accomplishments that we're proud of

We're proud to have been able to gain some experience with machine learning because our entire team had never worked with this before.

What we learned

Some members of the group learned how to use HTML and CSS for the front end. Other members learned Keras for the back end.

What's next for Real Estate Predictor

We'd like to integrate the Sentiment Analyzer, the Machine learning algorithm, and front end into one cohesive project that an individual would be able to use easily by simply entering in a headline and getting the change in housing price index as an output all on one page. To improve its accuracy in predicting the housing price index changes we could feed the program a bigger set of data.