| NEWS | R Documentation |
Necessary changes were enacted so that the package can be built within RStudio, e.g., roxygen comments were added to R source files.
Created functions that create panel designs and that compute and
plot power for panel designs given a linear trend. Function
revisit_dsgn creates a panel revisit design. Function
revisit_bibd creates a balanced incomplete block panel revisit
design. Function revisit_rand creates a revisit design with
random assignment to panels and time periods. Function
panel_summary summarizes characteristics of a revisit panel
design. Function power.dsgn calculates power for multiple panel
designs. Function cov.panel.dsgn creates the covariance matrix
for a panel design. Function plot.powerpaneldesign plots power
curves for panel designs.
Modified C function readDbfFile to avoid PROTECT errors.
None.
Modified C functions insideAreaGridCell,
insideLinearGridCell, linSample, linSampleIRS,
numLevels, pickAreaSamplePoints,
pickLinearSamplePoints, and pointInPolygonFile to avoid
PROTECT errors.
Modified function change.analysis to assign the support_2
variable from the design_2 data frame rather than from the design_1 data
frame.
Modified function change.analysis to use a revised procedure
for ensuring that each repeated visit site is present in both surveys
for subpopulation estimation.
Modified C functions writeShapeFilePoint and
writeShapeFilePolygon to correct memory access errors.
Modified C function readDbfFile to ensure that missing values
(NA) are handled correctly.
Inserted a SystemRequirements field in the DESCRIPTION file and
modified functions grts, grtsarea, grtslin,
grtspts, irs, irsarea, irslin,
read.dbf, and read.shape to prevent the functions from
being executed on big-endian processors.
Removed the C header file named order.h and replaced C
functions readBigEndian and readLittleEndian with the
version of those functions from spsurvey version 3.1.
Created a vignette for change estimation.
Modified functions change.analysis and change.est to
include estimation for difference in median values for change estimation
using continuous variables.
Created a function named examine that examines variables in a
data frame by printing either a table or the results of a call to the
describe function in the Hmisc package
Modified function sp2shape to accommodate objects of class
"SpatialDesign" created by either the grts or irs functions.
Added a C header file named order.h that determines whether
the CPU employs big-endian or little-endian byte order. Also, modified
C functions readBigEndian and readLittleEndian to use
constants created by the header file to ensure that computer words are
read in the correct byte order.
Modified function input.check to ensure that use of numeric
variables for arguments stratum and cluster are
handled correctly when finite or continuous population correction
factors are utilized.
Modified function cdf.test to ensure upper bounds that define
classes for the CDFs are unique. Also, modified functions
cont.cdftest and cdf.test to reduce the minimum number of
upper bounds that define classes for the CDFs from three to two.
Modified functions sp2shape and read.dbf to ensure
that missing values are handled without error for character vectors.
Modified functions grts and irs to ensure that a data
frame object is assigned to argument att.frame when argument src.frame
is assigned value "att.frame".
Modified function change.analysis to ensure that each
repeated visit site is present in both surveys for subpopulation
estimation.
Modified function read.sas to use
"C:/Program Files/SASHome/SASFoundation/9.4/sas.exe" as the default
value for argument sascmd.
Modified vignettes to use data sets from the data directory for the package.
Modified C functions to ensure that variables passed to function malloc are of type unsigned integer.
Created a class named SpatialDesign that contains class
SpatialPointsDataFrame, which is defined in package sp. The class is
used for the output objects created by functions grts and
irs. Objects of class SpatialDesign include the survey design
list as a slot rather than as an attribute, which was used previously.
Created S3 and S4 methods for generic functions plot and summary for objects of class SpatialDesign.
Modified C functions to change the type declaration for variable
shpFileName from "const char *" to "char * restrict".
Modified C functions writeDbfFile,
writeShapeFilePoint, and writeShapeFilePolygon to change
the return type from void to SEXP NULL, i.e., R_NilValue.
Modified function input.check to remove missing values from
final adjusted weights prior to checking that all values are positive.
Modified functions grtspts, grtslin, and
grtsarea to ensure that argument nlev (number of hierarchical
levels) for C fucntion constructAddr was stored as type integer.
Added a cex.main argument to the cont.cdfplot and
cdf.plot functions.
Modified function warnprnt to use correct variable names in
the output data frame so that partial matching warnings for the names are
not generated.
Created a data directory that contain rda versions of the data files used by vignettes.
Modified function change.analysis to allow analysis of
surveys with no repeat visit sites.
Modified function change.est to ensure that levels for a
categorical response variable are listed in the desired order rather than
sorted order.
Modified function localmean.weight to allow recovery from
an error in the singular value decomposition function (La.svd) that is
called by the generalized inverse function (ginv) in the MASS package.
Also, created a support function named localmean.weight2.
Modified function grtspts and C function numLevels to
calculate the sampling grid random shift values only once rather than each
time the number of hierarchical levels is incremented.
Modified functions attrisk.analysis, cat.analysis,
change.analysis, cont.analysis, and relrisk.analysis
to replace NA values with FALSE in the indicator variables for
subpopulations.
Modified function spbalance to include all grid cells with
nonzero extent for the frame when calculating the Pielou evenness measure
for a rectangular grid.
Modified function localmean.weight to ensure that the initial
set of weights are inversely proportional to distance.
Modified functions cont.cdftest and cdf.test to ensure
that the number of classes into which the CDFs will be divided (binned) is
at least three.
Modified function dframe.check to ensure proper handling of
repeated site ID values in the sites data frame.
Created a vignette that presents deconvolution of a cumulative distribution function (CDF) generated by a GRTS survey design.
Created a function named spbalance that calculates spatial
balance metrics for a survey design.
Created functions named sbcframe and sbcsamp that
calculate spatial balance grid cell extent and proportions for a sample
frame and for a survey design, respectively.
Modified function change.est to include calculation of
resource size change estimates for categorical variables.
Created a function named changevar.size to calculate
covariance or correlation estimates of the estimated change in class
resource size estimates between two probability surveys
Modified function change.est to correct errors in the output
values for categorical variables from survey designs that lack repeat
visit sites.
Modified the following functions to assign consecutive numbers to the row names for the output data frame(s): attrisk.analysis, cat.analysis, change.analysis, cont.analysis, cont.cdftest, and relrisk.analysis.
Created a function named change.analysis that conducts change
analysis for a collection of response variables (categorical and
continuous) generated by two probability surveys.
Created functions named change.est, changevar.prop,
and changevar.mean to calculate estimates of change and its
associated variance.
Created a vignette that presents cumulative distribution function (CDF) analysis of a GRTS survey designs.
Added a function named ash1.wgt that computes the average shifted histogram (ASH) for weighted data.
Created an .onAttach function that prints a message when the spsurvey package is attached. Removed the .onLoad function, which prior versions used to to print a startup message.
Modified function grtspts to correct an error that occurs
when argument src.frame equals "att.frame" and the number of
hierarchical levels equals eleven.
Modified C functions printAddrList, printColCharList,
and printDbf to replace calls to the printf function with calls to
the Rprintf function.
Modified C function pickAreaSamplePoints to correct an error
that occurs when selecting sample points for PolygonZ and PolygonM type
shapefiles.
In order to reduce package size, removed demonstration (demo) R scripts and the associated data directory.
Created vignettes that present analyses of GRTS survey designs for finite, linear, and areal resources.
Created a function named attrisk.analysis that conducts
attributable risk analysis for a collection of response variables
generated by a probability survey.
Created functions named attrisk.est and attrisk.var to
calculate estimates of attributable risk and its associated variance.
Modified function relrisk.est to change the way the relative
risk estimate is calculated for a stratified sample.
Modified functions cat.analysis, cont.analysis,
cont.cdftest, relrisk.analysis, category.est,
cdf.est, cdf.decon, total.est, cdf.test,
relrisk.est, catvar.prop, catvar.size,
cdfvar.prop, cdfvar.total, cdfvar.size.prop,
cdfvar.size.total, dcdfvar.prop, dcdfvar.total,
dcdfvar.size.prop, dcdfvar.size.total, total.var,
cdfvar.test, and relrisk.var to allow variance estimates to
be calculated when a two stage sample has stage one sampling units that
contain a single stage two sampling unit. Variance for those stage one
sampling units is calculated using the mean of the variance estimates for
stage one sampling units that contain two or more stage two sampling
units.
Modified function grtslin to improve efficiency of sample
point selection. Created a C function named insideLinearGridCell
that, for each grid cell selected to receive a sample point, returns the
ID value for shapefile records contained in the cell and the clipped
length of the polyline segments within the cell for each record. Created
a C function named pickLinearSamplePoints that selects sample
points.
Modified function grtsarea to improve efficiency of sample
point selection. Created a C function named insideAreaGridCell
that, for each grid cell selected to receive a sample point, returns the
ID value for shapefile records contained in the cell and the clipped area
of the polygon within the cell for each record. Created a C function
named pickAreaSamplePoints that selects sample points.
Modified function input.check to include determination of
whether a valid value was provided for argument vartype.
Modified functions grts and irs so that correct survey
design weights are created for linear and areal designs when the type of
random selection is "continuous".
Modified functions grtspts, grtslin, and
grtsarea to execute without error for a stratum sample size of one.
Modified function dframe.check to terminate execution when
missing values are encountered in the logical variable of the sites data
frame.
Created vignettes that present examples of GRTS survey designs for finite, linear, and areal resources.
Created a function named geodalbers that projects latitude
and longitude (spheroid) models of the globe to Albers projection in the
plane.
Modified function dsgnsum to produce summary tables for the
actual set of design variables that are present rather than a standard set
of design variables.
Modified function dframe.check to terminate execution when
missing site ID values or repeated site ID values are encountered in the
sites, design, subpop, data.cat,
data.cont, or data.rr data frames. Note that
dframe.check is called by functions cat.analysis,
cont.analysis, cont.cdftest, and relrisk.analysis.
Modified function cont.cdftest to correct an error caused by
improper handling of an empty subpopulation.
Modified function relrisk.analysis to correct an error
resulting from inclusion of more than one response variable in the
response.var argument.
Eliminated use of argument unitsize (known sum of
size-weights) by package functions. Restricted argument popsize to
provide only the known size of the resource for use in ratio adjustment of
estimators. Created a new argument named pcfsize to provide
resource size for calculation of finite and continuous population
correction factors for single-stage samples.
Modified functions cat.analysis, cont.analysis,
cont.cdftest, and relrisk.analysis to add logical variables
to their argument lists that specify use of finite or continuous
population correction factors or use of size-weights in analysis routines.
Modified functions category.est, cdf.decon,
cdf.est, cdf.test, relrisk.est and total.est
to add logical variables to their argument lists that specify use of
finite or continuous population correction factors or use of size-weights.
Modified function irs to eliminate checking for existence of
x-coordinates and y-coordinates when the type of frame is "finite" and the
frame is included in the att.frame data frame.
Modified C functions parseFields and readDbfFile to
ensure that blank values in the shapefile attributes (dbf) file are
converted to R missing values (NA).
Modified C function writeDbfFile to output blank values for R
missing values (NA) when creating the shapefile attributes (dbf) file.
Modified functions grts, irs, and sp2shape to
ensure that an output shapefile attributes (dbf) file containing character
variables with missing values can be read without error.
Created a function named cont.cdftest that tests for
differences between cumulative distribution functions (CDFs) for pairs of
subpopulations within a population Type for a collection of
response variables generated by a probability survey. Also, modified
function cdf.test so that it is consistent with functions that are
called by high-level functions cat.analysis and
cont.analysis.
Modified function grtspts so that, when source of the frame
equals "att.frame", the current number of grid levels for hierarchical
randomization and the final number of grid levels is printed to the
console while the function is executing.
Created a function named relrisk.analysis that conducts
relative risk analysis for a collection of response variables generated by
a probability survey. Also, renamed function relrisk to
relrisk.est and modified the function so that it is consistent with
functions that are called by high-level functions cat.analysis and
cont.analysis.
Modified function grtspts and C function numLevels to
terminate the algorithm for determining the number of grid levels for
hierarchical randomization when the maximum value of total inclusion
probability for the grid cells stops changing as the number of levels
increases.
Created functions named cdf.plot and cont.cdfplot that
create cumulative distribution function (CDF) plots using the data frame
named "CDF" contained in the output object created by function
cont.analysis. Function cdf.plot creates a single CDF plot,
and function cont.cdfplot creates a set of CDF plots.
Created a function named read.sas that can read SAS datasets
or a SAS XPORT (transport) file.
Modified C functions intersect, linSampleIRS,
lintFcn, and linSample to eliminate warning messages that
were generated during package creation.
Modified function grts so that argument do.sample
provides a value for each stratum.
Modified C functions getRecordShapeSizes, readDbfFile,
parsePoints, parsePointsZ, parsePointsM,
parsePolygon, parsePolygonZ, and parsePolygonM to
generate error messages and terminate execution when a shapefile
containing a Null record is encountered.
Modified functions irslin and irsarea in addition to C
function getRecordIDs to ensure that sample points are selected in
random order for linear and areal IRS designs.
Modified function grts to ensure that, when the type of
random selection is "unequal", an oversample is apportioned correctly
whenever the category sample sizes are proportional to the oversample
size.
For C functions that read shapefiles from the current directory,
replaced calls to _findfirst and _findnext with code using
calls to readdir. These changes were implemented to facilitate
portability of the package.
Created a C function named matchFiles that determines whether
file names in the current directory have a desired file extension. This
change was implemented to facilitate portability of the package.
Modified functions grts and irs to accommodate use of
a factor for the ID variable in the attributes data frame.
Modified functions grts and irs to ensure that sample
weights are correctly adjusted when an oversample is present and the type
of random selection is "Continuous".
None.
Modified C functions getRecordShapeSizes and lintFcn
to accommodate Polyline shapefiles that have multiple parts.
Modified functions dsgnsum and sp2shape to accommodate
the change in representation from AttributeList to
data.frame for the data slot of sp package objects belonging to
class SpatialPointsDataFrame.
Modified functions grts and irs to print a warning
message when the type of frame equals "finite" and a stratum name in the
design list matches only a single value in the stratum column of the
attributes data frame. For this case, function grtspts or
irspts is not called since the sample will be composed of a single
point.
Modified functions grts, grtspts, grtslin, and
grtsarea to change the maximum value allowed for arguments startlev
(the initial number of hierarchical levels to use for the GRTS grid) and
maxlev (the maximum number of hierarchical levels to use for the GRTS
grid) from 12 to 11.
Added an example polylines dataset to the data directory.
Modified functions grts and irs to allow use of an sp
package object as the source of the frame. An argument named
sp.object was added to the argument list for grts and
irs.
Modified functions grts, grtspts, grtslin,
grtsarea, irs, irspts, irslin, and
irsarea to remove use of argument xy.frame as an option for
source of the frame. Then modified functions grts, grtspts,
irs, and irspts to allow incorporation of frame coordinates
into the attributes data frame when the type of frame equals "finite".
Also, removed argument elmsize from functions grts and
irs since the argument no longer was required.
Modified functions grts and irs to print a warning
message when the type of frame equals "finite" and a stratum name in the
design list matches only a single value in the stratum column of the
attributes data frame. For this case, function grtspts or
irspts is not called since the sample will be composed of a single
point.
Modified functions grts and irs to ensure that the ID
values for elements of the frame provided in att.frame are unique.
Modified functions grts and irs to ensure that valid
values are provided for the panel and caty.n arguments in
the design list.
Added an example polygons dataset to the data directory.
Incorporated the CHANGES, README, and UserGuide files into the help page.
None.
Created a function named sp2shape and a C function named
writeShapeFilePolygon that convert objects created by package
sp to ESRI shapefiles. Also, renamed the C function writeShapeFile
to writeShapeFilePoint.
Modified function irsarea and created a C function
named getRecordIDs to ensure that an IRS sample is selected when
argument type.frame is set to "area" in function irs.
Modified the C functions so that the package can accommodate M-type shapefiles.
Modified functions grts and irs to ensure that the
levels of mdcaty (the variable in the attributes data frame that
specifies the unequal probability category for each element in the frame)
are equivalent to the names in caty.n (the vector of sample sizes
for each category in mdcaty, which is required for each element of
the design list for which the selection type is "Unequal").
Modified functions grts and irs to ensure that the
columns of xy.frame are named "x" and "y" when xy.frame is
provided and type.frame equals "finite".
Modified functions grts and irs so that the sample
weights are correctly adjusted when an oversample is requested and the
realized sample size is less than the desired sample size.
This is the original version of the package.
None.