Information about Brockwell and Davis Data Files

The files on the floppy that comes with Introdution to Time Series and
Forecasting by Peter J. Brockwell and Richard A. Davis (Springer 1996)
are plain text files designed to work with the program ITSM on the same
floppy.  Although many of them can be read "as is" by MacAnova using
vecread() or readcols, quite a number cannot because of descriptive
information that is on the same lines as the data.  For this reason, I
have created additional data files that can be used more easily with
MacAnova and can easily be edited to work with almost any statistics
program that can read text files.

1.  For each file on the floppy with name of the form foo.dat (for
example, airpass.dat), I have created a file of the form foo.txt (for
example, airpass.txt).  This has the data but with a header consisting
of descriptive lines starting with "#".  For example, the first few
lines of airpass.txt are

      # Data from file airpass.dat distributed with Introduction to
      # Time Series and Forecasting by Peter J. Brockwell and
      # Richard A. Davis, Springer 1996
      # (C) 1996 Springer-Verlag New York, Inc.
      # Monthly totals of international airline passengers
      # from January 1949 to December 1960 (same as Series G,
      # Box & Jenkins, P. 531)  N = 144
       112
       118
       132
       . . .

The following files have a single variable:
        airpass.txt  cdid.txt     lake.txt     signal.txt
        apph.txt     cdin.txt     lead.txt     sres.txt
        appi.txt     chaos.txt    lres.txt     strikes.txt
        appj.txt     deaths.txt   oshorts.txt  sunspots.txt
        appk.txt     dowj.txt     respc.txt
        arch.txt     ex611.txt    sales.txt

The following files have two variables in two columns:
        appjk2.txt   djaopc2.txt  goals.txt    lynx.txt
        djao2.txt    djaopcf.txt  ls2.txt

In a few cases I have created three variable files in three columns by
adding columns containing the year (always column 1) and quarter or
month (column 2) with the series itself in column 3.  These files are
        austres.txt  deaths.txt   gnfp.txt     wind.txt
        beer.txt     elec.txt     imports.txt
        chocs.txt    finserv.txt  polio.txt

You can read a single variable file like apph.txt by

  Cmd> y <- vecread("apph.txt", skip:"#")

As usual, on Windows, Macintosh or Motif, you can use "" as the file name.

You can read a file with two variables into variables x1 and x2 by, say,

  Cmd> readcols("djao2.txt", x1, x2, skip:"#")

You can read a file with year, month and data as, for example

  Cmd> readcols("beer.txt",year,month,beer,skip:"#")

These files ought to be readable by Minitab and by any program that
ignores lines that start with "#".  To read the data in other programs,
including ITSM, that can read data from text files, you will need to use
an editor or word processor to delete the lines starting with "#".

2.  File BDData.txt contains the data from all the files in the form of
data sets readable by matread().  The data set names just omit the
".dat" of the files on the diskette or the ".txt" of the new files.  For
example, you can read the data originally in djao2.dat and also in
djao2.txt by

  Cmd> djao2 <- matread("BDData.txt","djao2")

You can get a list of all the data sets in the file by

  Cmd> matread("BDData.txt","info")


C. Bingham, Fri Jan 15 07:49:46 CST 1999