Hedonic Pricing#

The hedonic pricing method constructs house price indices by modeling sale prices as a function of observable property characteristics (such as size, location, and age) and time. This approach uses all available transactions and can produce robust indices when the model is well-specified. It offers transparency and flexibility, making it a common baseline method for house price analysis.

Note

Background on basic model construction for hedonic pricing models can be found in:

Bourassa et al. (2006), “A Simple Alternative House Price Index Method”, Journal of Housing Economics, 15(1), 80-97. DOI: 10.1016/j.jhe.2006.03.001.

Data Preparation#

For hedonic pricing, your data should include:

  • A date column (e.g., “sale_date”)

  • A price column (e.g., “sale_price”)

  • Property characteristics columns (e.g., “sqft”, “bedrooms”, “bathrooms”)

  • A transaction identifier column (e.g., “sale_id”)

You can prepare your data as follows:

>>> from hpipy.datasets import load_ex_sales
>>> from hpipy.period_table import PeriodTable
>>> from hpipy.trans_data import HedonicTransactionData

# Load sales data.
>>> df = load_ex_sales()

# Create period table.
>>> sales_hdata = PeriodTable(df).create_period_table(
...     "sale_date",
...     periodicity="monthly",
... )

# Prepare hedonic data.
>>> trans_data = HedonicTransactionData(sales_hdata).create_transactions(
...     prop_id="pinx",
...     trans_id="sale_id",
...     price="sale_price",
... )

Creating the Index#

Once your data is prepared, you can create the hedonic price index:

>>> from hpipy.price_index import HedonicIndex

# Create the index.
>>> hpi = HedonicIndex.create_index(
...     trans_data=trans_data,
...     price="sale_price",
...     date="sale_date",
...     dep_var="price",
...     ind_var=["tot_sf", "beds", "baths"],
...     estimator="base",  # or "robust", "weighted"
...     log_dep=True,
...     smooth=True,
... )

Parameters#

The main parameters for hedonic index creation are:

Parameters

estimatorstr

The type of estimator to use:

  • “base”: Standard OLS estimation

  • “robust”: Robust regression (less sensitive to outliers)

  • “weighted”: Weighted regression

dep_varstr

Dependent variable to model.

ind_varlist

Independent variables to use in the model.

log_depbool

Whether to use log of price as dependent variable (recommended).

Advanced Usage#

For more control over the hedonic model:

>>> from hpipy.price_index import HedonicIndex
>>> from hpipy.price_model import HedonicModel

# Create and fit the model.
>>> model = HedonicModel(trans_data).fit(
...     dep_var="price",
...     ind_var=["tot_sf", "beds", "baths"],
...     estimator="base",
...     log_dep=True,
... )

# Create the index.
>>> hpi = HedonicIndex.from_model(model, trans_data=trans_data, smooth=True)

Feature Engineering#

The hedonic method often benefits from careful feature engineering:

  1. Numeric Transformations:

    >>> import numpy as np

# Log transform skewed features.
>>> df["log_sqft"] = np.log(df["tot_sf"])

# Create interaction terms.
>>> df["price_per_sqft"] = df["sale_price"] / df["tot_sf"]

  2. Categorical Features:

    >>> import pandas as pd

>>> cat_cols = ["use_type", "area"]

# One-hot encode categorical variables.
>>> df = pd.get_dummies(df, columns=cat_cols)

  3. Spatial Features:

    >>> lat_col, lon_col = "latitude", "longitude"

# Create location-based features.
>>> df["lat_lon"] = (
...     df.loc[:, [lat_col, lon_col]]
...     .round(2)
...     .astype(str)
...     .agg("_".join, axis=1)
... )

Evaluating the Index#

Evaluate the hedonic index using various metrics:

>>> from hpipy.utils.metrics import volatility
>>> from hpipy.utils.plotting import plot_index

# Calculate metrics.
>>> vol = volatility(hpi)

# Visualize the index.
>>> plot_index(hpi, smooth=True).properties(title="Hedonic Index")
alt.LayerChart(...)