Rsafd

Rsafd is an R package that accompanies René Carmona's book Statistical Analysis of Financial Data in R. It gathers the datasets and reusable functions referenced throughout the text so readers can reproduce examples, explore additional analyses, and adapt the techniques to their own financial time-series problems.

Package highlights

• Curated financial datasets spanning equity indexes, interest rates, energy markets, bond curves, and alternative assets (e.g., DSP, SPFUT, FRWRD, us.bis.yield, bitcoin).
• Time-series utilities for filtering, decomposition, forecasting, and volatility modeling (kalman, sstl, pred.ar, garch, sim.garch).
• Extremes and heavy tails with generalized extreme value and generalized Pareto distributions, L-moments, tail plots, and POT estimation (gev.ml, fit.gpd, tailplot).
• Dependence modeling through copulas, multivariate normals, kernel density/regression, and correlation diagnostics (fit.copula, rcopula, dmvnorm, kreg).
• Option pricing and fixed-income helpers including Black–Scholes, implied volatility, Nelson–Siegel forward curves, and coupon bond analytics (bscall, isig, fns, bns).

A complete index of exported symbols is available in INDEX, while man/ hosts the Rd documentation files used by R's help system.

Installation

# Install dependencies first
install.packages(c(
  "glasso", "MASS", "plot3D", "qgraph", "quadprog", "quantreg",
  "robustbase", "scatterplot3d", "splines", "timeDate", "tseries"
))

# Install Rsafd from the latest GitHub release
if (.Platform$OS.type == "windows") {
  install.packages(
    "https://github.com/PrincetonUniversity/Rsafd/releases/latest/download/Rsafd.zip",
    repos = NULL,
    type = "source"
  )
} else {
  install.packages(
    "https://github.com/PrincetonUniversity/Rsafd/releases/latest/download/Rsafd.tar.gz",
    repos = NULL,
    type = "source"
  )
}

The package requires R 4.2.0 or newer.

Getting started

library(Rsafd)

# Load one of the bundled datasets
data("DSP")
plot(DSP, type = "l", main = "Daily S&P 500")

# Fit a GARCH(1,1) model to S&P 500 log returns
returns <- diff(log(DSP))[-1]
fit <- garch(returns, order = c(1, 1))
summary(fit)

# Estimate a generalized Pareto tail on extreme returns
library(quantreg)
exceedances <- returns[returns < quantile(returns, 0.95)]
fit_gpd <- fit.gpd(exceedances, threshold = quantile(exceedances, 0.9))
plot(fit_gpd)

See help(package = "Rsafd") for a searchable list of all datasets and helper functions, and consult the book for guided exercises that pair with each object.

Repository layout

• R/ contains the lazily-loaded compiled R code.
• data/ stores serialized datasets in R's .rda format.
• man/ holds Rd documentation consumed by ?foo help pages.
• html/ supplies the pre-built reference index for R CMD help.
• DESCRIPTION, NAMESPACE, LICENSE, and INDEX provide standard package metadata.

Building from source

R CMD build Rsafd
R CMD check Rsafd_1.0.1.tar.gz

Running R CMD check is recommended before submitting changes or distributing a new tarball.