ecoGroup: An example of a transformation 'function'

In henkelstone/NPEL.Classification: Classifaction and data-handling routines for NPEL Caribou Project

Description

This dataset is a list, cast in matrix format, that can be used to lump ecotypes into coarser groups for analysis and plotting. The current implementation includes vectors for: transformation, naming/labelling, and suggested plotting colours

Usage

ecoGroup[[which.grouping,type.of.data]]

Arguments

which.grouping	one of c('identity','domSpecies','domGroup','maxGranularity'), see Format.
type.of.data	one of c('transform','labels','colours'), see Format.

Format

A list cast into matrix format with four rows (grouping scenarios) and three columns (possible data types for that scenario). The four different grouping scenarios are:

• Identity (‘identity’) maps the class onto itself, that is, there is no lumping, but labels and colours are provided.

• Dominant Species (‘domSpecies’) groups classes by the dominant (tree) species.

• Dominant Group (‘domGroup’) groups classes by the dominant growth-form.

• Maximum Granularity (‘maxGranularity’) groups the classes in what is deemed reasonable for our data, that is, we attempt to maximize the number of groupings while maintaining a reasonable number of samples in each group for statistical significance and reliability.

The different types of data that can be extracted are:

• Transformation Vector (‘transform’) a transformation vector converting input factors (ecoSites) to output factors (groupings). Functionally it is a lookup table that maps input classes to output classes. See ‘examples’ for how to use it.

• Labels (‘labels’) meaningful names for the output groups.

• Plotting Colours (‘colours’) suggested plotting colours for the output groups. These are easy to pass to, for example, plotting functions.

Details

This built-in transformation 'function' encodes several different grouping scenarios. Each is provided with labels for the groups as well as suggested colours. Following is a description of the four build-in encoding for the FEC classification system McLaughlan, M. S., Robert A. Wright, and R. D. Jiricka. Field Guide to the Ecosites of Saskatchewan's Provincial Forests. Prince Albert, Sask: Ministry of Environment, 2010.

Ecotype	Dominant Species Class	Dominant Group Class	Maximum Granularity Class
BS 01 -- Sand heather	1 (barren)	1 (barren)	1 (sparse vegetation)
BS 02 -- Lichen felsenmeer	1	1	1
BS 03 -- Jack pine, blueberry, lichen	2 (pine)	2 (conifer)	2
BS 04 -- Jack pine, black spruce, feathermoss	2	2	3
BS 05 -- Jack pine, birch, feathermoss	2	2	4
BS 06 -- Jack pine, aspen, alder	2	2	5
BS 07 -- Black spruce, blueberry, lichen	3 (bs)	2	6
BS 08 -- Black spruce, birch, lichen	3	2	7
BS 09 -- Black spruce, pine, feathermoss	3	2	8
BS 10 -- Black spruce, birch, feathermoss	3	2	9
BS 11 -- White spruce, fir, feathermoss	4 (ws)	2	10
BS 12 -- White spruce, crowberry, feathermoss	4	2	11
BS 13 -- Birch, black spruce, aspen	5 (birch)	3 (decid)	12
BS 14 -- Birch, lingonberry, lab tea	5	3	13
BS 15 -- Aspen, birch, alder	6 (aspen)	3	14
BS 16 -- Black spruce, poplar, alder swamp	7 (swamp)	4 (wetland)	15 (group swamp in with treed bog)
BS 17 -- Black spruce bog	8 (bog)	4	15
BS 18 -- Lab tea shrubby bog	8	4	16
BS 19 -- Graminoid bog	8	4	17 (sparse bog)
BS 20 -- Open bog	8	4	17
BS 21 -- Tamarack fen	9 (fen)	4	18 (upright fen)
BS 22 -- Leatherleaf fen	9	4	18
BS 23 -- Willow shrubby fen	9	4	19
BS 24 -- Graminoid fen	9	4	20 (sparse fen)
BS 25 -- Open fen	9	4	20
BS 26 -- Rush sandy shore	10 (shore)	4	21 (shoreline)
BS 27 -- Sedge rocky shore	10	4	21

Warning

The following expressions will not work as transformation functions. See factor, especially the Warning section for more information on this gotcha.

siteData$Ecotype returns a factor list, not a numeric list representing the factors.

ecoGroup[['identity','transform']][siteData$Ecotype] will regroup based on the factor level indices, NOT the factors.

as.numeric(siteData$Ecotype] returns the factor indices, not a numeric representation of the factors.

(1:27)[siteData$Ecotype] returns the factor index rather than the ecoGroup.

Examples

data (siteData)
ecoGroup[['domSpecies','transform']][ factorValues(siteData$ecoType) ]

## Code for generating this object--so you can make your own!
# Create a shell to fill with data
ecoGroup <- list(); length(ecoGroup) <- 12;
dim(ecoGroup) <- c(4,3)               # Three rows (classifiers), three columns (type of data)
rownames(ecoGroup)=c('identity','domSpecies','maxGranularity','domGroup')
colnames(ecoGroup)=c('transform','labels','colours')

# Fill with data in increasing order of lumping
ecoGroup['identity',] <- list(
  c(1:27),
  paste0('BS',1:27),
  c('#FFFF00','#E6E600','#FF9900','#E47A07','#BD6A06','#965B05','#4C7300','#5E8D00','#70A800',
    '#82C300','#008C4B','#00A04B','#ABFF8F','#42FF07','#7EFF54','#41DBCF','#00A884','#3FC5A5',
    '#7EE2C6','#BEFFE8','#97E2FF','#8DD4F0','#83C6E1','#79B8D2', '#6FABC3','#AC44E7','#C077E3')
)
ecoGroup['maxGranularity',] <- list(
  c(1,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,15,16,17,17,18,18,19,20,20,21,21),
  c('Sparse',paste0('BS',3:15),'Treed bog/swamp','BS18','Sparse bog','Upright fen','BS23','Sparse fen','Shoreline'),
  c('#E6E600','#FF9900','#E47A07','#BD6A06','#965B05','#4C7300','#5E8D00','#70A800','#82C300',
    '#008C4B','#00A04B','#ABFF8F','#42FF07','#7EFF54','#00A884','#3FC5A5','#BEFFE8','#97E2FF',
    '#83C6E1','#6FABC3','#C077E3')
)
ecoGroup['domSpecies',] <- list(
  c(1,1,2,2,2,2,3,3,3,3,4,4,5,5,6,7,8,8,8,8,9,9,9,9,9,10,10),
  c('1'="Barren",'2'="Pine",'3'="Black Sp",'4'="White Sp",'5'="Birch",'6'="Aspen",'7'="Swamp",
    '8'="Bog",'9'="Fen",'10'="Shore"),
  c('E6E600','E47A07','70A800','00A04B','42FF07','41DBCF','3FC5A5','83C6E1','C077E3','004DA8','000000')
)
ecoGroup['domGroup',] <- list(
  c(1,1,2,2,2,2,2,2,2,2,2,2,3,3,3,4,4,4,4,4,4,4,4,4,4,4,4),
  c("Barren","Conifer","Decid","Wetland"),
  c('E6E600','00A04B','42FF07','83C6E1')
)

henkelstone/NPEL.Classification documentation built on May 17, 2019, 3:42 p.m.