User Guide#

Welcome to the hpiPy User Guide. This guide walks you through creating and evaluating House Price Indices using various methods—from repeat sales to hedonic pricing to machine learning.

Price Index Methods#

Use one of the following methods to build house price indices:

📈 Repeat Sales

Build house price indices by pairing repeat sales of unchanged properties.

Repeat Sales
🏘️ Hedonic Pricing

Model house price indices as a function of house features like size, location, and age.

Hedonic Pricing
🌲 Interpretable Random Forest

Use an ensemble of decision trees to learn complex, nonlinear price patterns.

Interpretable Random Forest
🧠 Bifurcated Neural Network

Apply a deep learning model to learn complex, nonlinear price patterns.

Bifurcated Neural Network

Evaluation & Comparison#

Once you’ve created your indices, use the following tools to evaluate and compare methods:

Accuracy Metrics

Measure how well an index predicts actual property values.

Accuracy Metrics
Volatility Metrics

Quantify the smoothness and stability of an index over time.

Volatility Metrics
Revision Metrics

Track how index values change with new data over time.

Revision Metrics
Series Metrics

Learn how to analyze and visualize index series.

Series Metrics
Comparative Analysis

Compare multiple index construction methods side by side.

Comparative Analysis