- New version (minor) including bug fixes, new functions, and function improvements.

- Improvements:
- Reviews and improves documentation.

- Deprecated functions:
`coordenadas()`

,`cdfTable()`

,`cdfStats()`

and`cdfPlot()`

. These functions were developed to fulfill the demands created by the PhD research project (2012-2016) and are no longer (or rarely) used. Some will be replaced, but all will move to a package currently under developed at https://github.com/samuel-rosa/ASRtools.

- Improvements:
- Reviews and improves documentation.

- Fixes:
- Deals with the fact that matrix objects now also inherit from class “array”.
- Deals with the fact that the suggested package
**geoR**has been orphaned on 2020-01-12. This is done by keeping the package in`Suggests`

and using it conditionally by employing`requireNamespace()`

.

- Improvements:
`plotCor`

: reverses default color ramp so that positive and negative correlations are depicted using red and blue colors, respectively.

- Improvements:
`plotHD`

: adds grid lines to histogram.`plotESDA`

: choose the position of the legend of the bubble plot

- Improvements on
`plotCor`

to deal with long column names in the correlation matrix. Long column names (> 5 characters) are now replaced with x1, x2, …, xn, where n is the number of columns in the correlation matrix. A message is issues identifying columns and their respective index.

- Adds
`plotCor`

, a function to create correlation plots using a colorblind-friendly palette.

- Evaluates the upcoming new version (3.0-0) of the
**car**package.

- Improved selection of the location to place the legend in
`plotHD`

.

- A new guess of
`p_sill`

is used when the variance is spatially uncorrelated. It is defined as 1e-3 times the guess of`nugget`

(before the`p_sill`

was set to 1e-12, which caused errors in`optim`

and`nlminb`

).

- FIX: corrected error in
`vgmICP`

caused by the recent update of package.*georob* `vgmICP`

: Improved estimation of the scale parameter (range) for monotonous variograms by using function`practicalRange`

from package.*geoR*

`vgmICP`

: Improve the guess of the partial sill for a pure nugget effect model.

- FIX: the definition of the lag-distance classes passed to
`georob::sample.variogram`

in`vgmLags`

was incorrect due to an error in the documentation the former. Instead of “a numeric vector with the upper bounds of a set of contiguous bins”, the argument`lag.dist.def`

of`georob::sample.variogram`

requires “a numeric vector with the*lower and*upper bounds of a set of contiguous bins”. The author and maintainer of the-package has been warned about this documentation error.*georob* - Improve guess of nugget variance by giving larger weight to the first lag because we now consider a minimum number of point-pairs per lag-distance class.

- FIX: the computation of the number of point-pairs per lag-distance class in
`vgmLags`

was incorrect because it neglected the fact that, in a full distance matrix, two points*a*and*b*form two pairs, i.e.*ab*and*ba*. The mistake is due to the fact that we use`SpatialTools::dist1`

to compute the distance matrix instead of`stats::dist`

.

- New functions have been added. They are designed for three purposes:
- Variogram estimation and analysis
- Evaluation of geostatistical models of uncertainty
- Calibrating random forest models

- New function
`vgmICP`

to guess the initial covariance parameters when fitting a variogram model. - Included description of funding agencies in package description.

- Improved documentation of the
`vgmLags`

-function, and added functionality to count the number of points or point-pairs per lag-distance class so that the user can decide if the returned lag-distance classes meet his/her needs. - New function
`vgmSCV`

to compute the proportion of the variance that is spatially correlated. - Minor improvement of the documentation of almost all functions.

- Developed a new function
`checkGMU`

which can be used to check the quality of a geostatistical model on uncertainty.

- New functions:
- Iteratively de-bias random forest regression.

- Packaging:
- Now using Travis-CI to check the package after every commit;
- Functions from default packages other than
`base`

are now imported to comply with the new change to the CRAN policy described at http://developer.r-project.org/blosxom.cgi/R-devel/NEWS/2015/06/29#n2015-06-29; - Most packages that used to be under the
`Imports:`

field in the`DESCRIPTION`

file now are under the`Suggests:`

field;

- Existing functions:
`cramer()`

:`NA`

s are now removed when computing the chi-squared statistic;`plotHD()`

: if the Box-Cox transformation is required, the data is now first checked for zeros and negative values. If zeros and negative values are present, the data is transformed into a positive-valued data, with the minimum value equal to 1.`plotESDA()`

: new argument to define the lag-distance classes;

- New functions:
- Evaluation of the data type contained in an object;
- Extract spatial trend data from an object of class
`likfit`

(see the geoR package for more info):`trend.terms()`

and`trend.matrix()`

; - Stratification and categorization of continuous data:
`breakPoints()`

and`stratify()`

; - Computation of lag-distance classes for variogram estimation:
`vgmLags()`

.

- Better documentation for several functions.
`cramer()`

: the cross-tabulation is now computed using`table()`

instead of`bigtabulate()`

to avoid dependency problems in Windows.

- Functions to evaluate the data type contained in an object.

- Migrated from R-Forge to GitHub;

- Test for objects of type
`"numeric"`

and`"integer`

.

- Functions used to optimize sample patterns were moved to the new package
**spsann**:`spJitter()`

,`spSANN()`

,`spsannMSSD()`

,`spsannPPL()`

, and`spsannCLHS()`

.

- Function to compute the Cramer’s V, a measure of association between categorical variables.

- Documentation is now generated using roxygen2.

- Started the fourth season of the initial development of the package. Many functions are being implemented in C++ to speed-up the computation.

- Improving the spatial simulated annealing.

- Started the third season of the initial development of the package. Significant changes were made in functions
`spSANN()`

,`spJitter()`

, and those used as objective functions in spatial simulated annealing.

- Spatial simulated annealing is now available:
`spSANN()`

.

- Corrected bugs in functions
`objPairs()`

,`objPoints()`

,`pairs_per_lag()`

,`points_per_lag()`

and`spJitter()`

. Added new criterion to`objPairs()`

and`objPoints()`

. Added new method to`spJitter()`

(finite set of candidate locations). Most of the changes speed up the functions and increase their functionalities.

- Included functions
`objPairs()`

,`objPoints()`

,`pairs_per_lag()`

and`points_per_lag()`

. These functions were designed to be used with spatial simulated annealing.

- Included functions
`spJitter()`

and`bbox2sp()`

.

- Started the second season of the initial development of the package. Included functions
`plotHD()`

and`plotESDA()`

.

- Removed
`toar()`

and`readsat()`

because they are better implemented using GRASS GIS commands directly.

- Second public release (R-Forge).
- Included function
`buildMS()`

, used to build a series of linear models using automated variable selection. Included function`statsMS()`

, used to obtain performance statistics of a series of linear models. Included function`plotMS()`

, used to plot the statistics of a models series. Corrected bugs detected during the first built.

- First public release (R-Forge).
- Several corrections in the documentation, preparing for the first public release.

- Seventh internal release.
- Function
`cdfTable()`

: added argument to return a data.frame object.

- Sixth internal release.
- Function
`cdfPlot()`

: added argument to define if the confidence limits of the CDF should be included in the plot.

- Fifth internal release.
- Function
`cdfPlot()`

now has options to show CDF parameters.

- Fourth internal release.
- Included function
`stepVIF()`

. Renamed many of the functions to comply with the style guide. The package was also renamed to**pedometrics**.

- Third internal release.
- Included functions
`readsat()`

and`toar()`

.

- Second internal release.
- Minor correction in function
`deltagcp()`

.

- First internal release.
- Six functions available:
`cdfplot()`

,`cdfstats()`

,`cdftable()`

,`coordenadas()`

,`deltagcp()`

, and`gcpvector()`

.