REPS (Real Estate Price Statistics) is an R package for computing price indices using a variety of hedonic regression and multilateral methods. It provides a unified interface to calculate quality-adjusted price indices across time, supporting classical index formulas as well as advanced techniques tailored to property price analysis. While developed with real estate in mind, the package is applicable to any domain where reliable, quality-adjusted price comparisons over time are needed. REPS was developed by Statistics Netherlands (CBS) with funding from Eurostat, aligning with international guidelines for compiling residential property price indices.
You can install the development version of REPS from GitHub using devtools:
# install.packages("devtools") # if devtools is not installed
::install_github("vivekag7/REPS") devtools
REPS offers a single entry point
calculate_price_index()
to compute price indices using
different methodologies. By changing the method
argument,
you can generate indices with any of the supported approaches, all from
the same dataset and function call. Supported index calculation methods
include:
All methods are accessed via calculate_price_index()
with a consistent interface. The function returns either a single index
or, for multiple methods, a list of results. A reference period can be
specified to normalize index values (default sets the first period to
100).
calculate_regression_diagnostics()
fits a separate
log-linear model per period and checks normality, R², autocorrelation,
and heteroscedasticity of residuals.
plot_regression_diagnostics()
visualizes these over time in
a 3×2 grid.plot_price_index()
plots one or more index series as a
time-series chart for comparison and exploration.data_constraxion
is
included (7,800 × 6), simulating real estate transactions. Use
data("data_constraxion")
to load it.vignette("calculate_price_index", package="REPS")
and
vignette("calculate_regression_diagnostics", package="REPS")
for step-by-step guides using the example dataset.REPS provides a comprehensive and extensible framework to compute and evaluate real estate price indices. Its flexible architecture enables comparison across methods and statistical validation of underlying models, suitable for both official statistics and research applications.