As graduating seniors, our class is looking to start our new chapter in life in different areas of the world. Our goal was to simplify the selection process in an efficient and smart way. We used a data set provided by CoStar from

What it does

Using a Machine Learning algorithm, we are able to predict the prices of an apartment in a specific area (DC Metro Area) based on specific criteria. Using our model, we can find the best value of an apartment listed.

How we built it

We created a front end visualization to represent the data in the data set based on geolocation. This visualization allows users to interact with each data point in order to see the minimum and maximum rent prices of each listing. In the back end, we pre-processed the data set and used dense neural networks.

Challenges we ran into

Making the prediction model more accurate after training it with our data set with TensorFlow Switched to ensemble learning after running dense neural networks Integrating the visualization graph in order to fit our data set

Accomplishments that we're proud of

Creating an integrated front end to back end application in the span of the hackathon.

What we learned

How to use ensemble machine learning to find the most accurate predictive model How to use use Cytoscape, a javascript library How to use d3.js

What's next for ParAppraise

Integrating more comparison options so we can compare different listings on our visualization Just-in-time visualization with new listings


Share this project: