Predicting Median Housing Prices by State in the USA Using Machine Learning

Overview

This project leverages machine learning techniques to predict median housing prices across different states in the USA over the next three years. The model utilizes various macroeconomic indicators such as interest rates, inflation, and unemployment rates to provide accurate forecasts. The aim is to assist real estate investors, policymakers, and financial institutions in making informed decisions by providing reliable predictions of housing market trends.

Project Structure

1. Data

• data/: Contains all datasets used in the project.
  • cpi.xlsx: Consumer Price Index data.
  • fed_interest_rate.csv: Federal interest rates.
  • house_rental_index.csv: House rental index data.
  • house_value_index.csv: House value index data.
  • pop_unemployment.xlsx: Population and unemployment data.
  • data.csv: Joined and cleansed dataset combining all the above tables.

2. Code

• us-housing-price.ipynb: Jupyter notebook containing all the code for data wrangling, preprocessing, EDA, model training, and prediction.

3. Documentation

• docs/: Contains all project documentation.
  • Predicting_USA_Median_Housing_Prices.pdf: The detailed white paper.
  • Presentation_Slides.pptx: PowerPoint presentation of the project.

Installation

Clone the repository

git clone https://github.com/yourusername/housing-price-prediction.git
cd housing-price-prediction

Create a virtual environment and install dependencies

python3 -m venv housing-price-venv
source housing-price-venv/bin/activate
pip install -r requirements.txt

Usage

Run the Jupyter Notebook

jupyter notebook us-housing-price.ipynb

This notebook includes:

• Data wrangling and preprocessing
• Exploratory Data Analysis (EDA)
• Model training and evaluation
• Predictions for future median housing prices

Project Highlights

Executive Summary

This project develops a machine learning model to predict median housing prices across US states, leveraging economic indicators, demographic data, and historical housing prices.

Business Problem

Predicting housing prices by focusing on macroeconomic indicators to help stakeholders make informed decisions.

Data Explanation

• Sources: Consumer Price Index, Federal Reserve Economic Data, Zillow, public housing databases, U.S. Census Bureau.
• Preparation: Handling missing values, scaling, and encoding.

Methods

• Models Used: Linear Regression, Random Forest, XGBoost.
• Evaluation Metrics: R-squared, MAE, RMSE.

Analysis and Results

• Feature Importance: Key features include rent, CPI index, total population, interest rates, and inflation.
• Model Performance: High accuracy with R-squared of 0.9988, MAE of 1817, and RMSE of 4093.

Future Work

• Real-Time Analysis: Integrate with real-time data feeds.
• Local Predictions: Extend to city or neighborhood levels.
• Policy and Planning: Aid in sustainable development.

Ethical Considerations

• Data Privacy: Ensure compliance with regulations.
• Bias and Fairness: Regular audits to prevent bias.
• Transparency: Use interpretable models.

Contributors

Jubyung Ha - jyubaeng@gmail.com