It's Free Real Estate

Here is an example of a result where a user is denied for a home purchase
Here is an example of a result where a user is approved for a home purchase

Inspiration

I am interested in Data Science and have never done a project related to FinTech. This challenge was an opportunity for me to just tackle a bunch of data head-on and find my own insights.

What it does

In this web app, users learn about the basics of home buying and can enter their own information in order to see if they currently qualify to buy a house. If they aren't, I give tips on how to improve their future chances.

How we built it

For data processing and analysis, I used Python. More specifically, I used Pandas to store and manipulate the data. To visualize it, I used Matplotlib. To train a machine learning model, I used SciKit Learn. For the web app, I used Streamlit to host my data and create the UI.

Challenges we ran into

My biggest challenge was figuring out what dependencies I needed. I had to have a certain version of streamlit and pandas in order for them to work together, so it took me a lot of time to figure that out. I also wanted to make the web app as accessible as possible, so I ran into challenges in designing something clean and approachable.

Accomplishments that we're proud of

I'm really proud of being able to take a bunch of data and turn it into something a person can actually use. Being able to apply each step of the data science pipeline is always so fulfilling!

What we learned

While I have used Streamlit before, I never connected a dataset to it and did in-app data analysis. I learned a lot about how to properly store data (to avoid storage issues). I also learned a lot about web design and what is the best way to represent information. I also was able to learn more about Logistic Regression.

What's next for It's Free Real Estate

There's so much I want to add in the next version! I would love to incorporate HTML/CSS to make the UI more appealing. I would also love to test other ML models (such as decision trees) to see which one best fits the data. In addition, I would love to optimize the current model by finetuning hyperparameters through a GridSearch.