Function to impose an change impact effect

Usage

genChangeData(
  pct.change,
  model,
  data,
  panels = NULL,
  eventsite.bnd = NULL,
  noneventcells = NULL,
  modelfits = NULL
)

Arguments

pct.change: percentage change to occur. e.g. 50% gives 50% site wide decline. If a redistribution effect (one of eventsite.bnd or noneventcells must be specified) then the change refers only to the event site cells. For an additional site-wide change when using a re-distribution, pct.change is a vector with the first number the event site re-distribution and the second, the site wide change.
model: A glm or gamMRSea model object. If specified, the fitted values from the model will be used to impose the change. Leave out if parameter modelfits is specified instead
data: Data frame used to fit model
panels: Character vector denoting the column of data containing the panel structure.
eventsite.bnd: A data frame containing the coordinates of a polygon defining the region of the defined event The variable names must match the coordinate system in the data.
noneventcells: logical (0/1) vector of length the same as the data indicating which cells the event did not occur (1).
modelfits: Character vector denoting the name of column in data containing the model fits.

Value

Returns a data frame twice the size of the original input with additional columns for panel id, eventphase and truth. The truth column represents the input data for the first half (pre-event; eventphase==0) and the second half has the change imposed (post-event; eventphase==1).

Author

Lindesay Scott-Hayward, University of St Andrews

Examples

data(nystedA_slim)
nysted$panelid <- paste0(nysted$yearmonth, ".t", nysted$TNO_ID)

initialModel<-MRSea::gamMRSea(response ~ 1 + as.factor(yearmonth)+depth +
                     x.pos + y.pos + offset(log(area)),  data=nysted,
                     family=quasipoisson)

# generate a 20% site-wide change:
changedata1<-genChangeData(pct.change=80, model=initialModel, data=nysted, panels="panelid")

# generate a 50% decline in the windfarm footprint:
data("nysted.bndwf")
changedata2<-genChangeData(pct.change=50, model=initialModel, data=nysted, panels="panelid",
                            eventsite.bnd=nysted.bndwf)
#> Loading required package: splancs
#> Loading required package: sp
#> 
#> Spatial Point Pattern Analysis Code in S-Plus
#>  
#>  Version 2 - Spatial and Space-Time analysis
#> Warning: List/data.frame components should be named x and y
#> Warning: List/data.frame components should be named x and y