House Improvement ROI Prediction Model

Group 4

James Siefert, Liam Hsieh, Panawannage Fernando, and Christopher Moussa

---

Background

The housing market plays a crucial role in today’s national economy, representing one of the most significant sectors in terms of financial transactions and investment opportunities. The valuation and prediction of housing prices have garnered significant attention from researchers, policymakers, and real estate professionals, as it directly impacts numerous stakeholders, including homeowners, buyers, sellers, investors, and government institutions.

Being able to predict housing prices based on a number of factors using historical data can effectively capture intricate relationships and identify crucical factors influencing housing prices. In this project, we separate and define these factors into two categories: changeable features (square footage, number of bedrooms and bathrooms, number of stories, presence of a guest room, central air conditioning, and furnishing status) and non-changeable features (proximity to a main road, presence of a basement, hot water heating, parking spaces, and preferred area status). These factors, when analyzed, can help buyers make informed decisions.

Implementation

This project looks to investigate the value increase associated with a number of home improvement projects to help homeowners maximize their return on investment (ROI). It uses a Kaggle dataset along with a number of trusted articles to guage values and other necessary information regarding home features and their relation to overall home price.

This project is broken into multiple phases: module creation, model training, model evaluation and refinement, and ROI calculation and integration. Each phase builds off of the previous in the Jupyter notebook to describe the machine learning process for the housing dataset, extensions of the data provided in the dataset through the creation of Python modules, and web appplication interfaces that link the trained model with the Python modules.

Goals

The goal of this project is to extend beyond the basic machine learning process with sample data and challenge ourselves to think about an extensible problem. While this project takes a number of steps past creating a linear regression model for the Kaggle dataset, it can be extended and advanced with the addition of more parameters and more data and training.

How to Run the Web UI locally

• Step 1: Go to web_app/ -> web_app.py
• Step 2: Run web_app.py
• Step 3: Go to http://127.0.0.1:5000 in your web browser
• Step 4: Alternate: Go into the project.ipynb file to view the application after web_app.py is running
• Step 5: Alternate 2: Contact James Siefert ([email protected]) and I can run a local server and send a URL for easy access.

How to Run the Web UI locally in a virtual environment in Windows and Python 3.10.9

• Step 1: open command prompt ("cmd" in search bar)
• Step 2: ensure you are using 3.10.9 by running python --version
• Step 3: navigate to project directory with cd <project folder path>
• Step 4: run python -m venv venv
• Step 5: run venv\Scripts\activate
• Step 6: install requirements: pip install -r requirements.txt
• Step 7: change directories to the web_app/ directory: cd web_app
• Step 8: run the web app: python web_app.py
• Step 9: Go to http://127.0.0.1:5000 in your web browser