Simulate data for developing project prioritizations for a priority threat management exercise (Carwardine et al. 2019). Here, data are simulated for a pre-specified number of features, actions, and projects. Features can benefit from multiple projects, and different projects can share actions.
simulate_ptm_data(
number_projects,
number_actions,
number_features,
cost_mean = 100,
cost_sd = 5,
success_min_probability = 0.7,
success_max_probability = 0.99,
funded_min_persistence_probability = 0.5,
funded_max_persistence_probability = 0.9,
baseline_min_persistence_probability = 0.01,
baseline_max_persistence_probability = 0.4,
locked_in_proportion = 0,
locked_out_proportion = 0
)numeric number of projects. Note that this
does not include the baseline project.
numeric number of actions. Note that this
does not include the baseline action.
numeric number of features.
numeric average cost for the actions. Defaults to
100.
numeric standard deviation in action costs. Defaults
to 5.
numeric minimum probability of the
projects succeeding if they are funded. Defaults to 0.7.
numeric maximum probability of the
projects succeeding if they are funded. Defaults to 0.99.
numeric minimum probability
of the features persisting if projects are funded and successful.
Defaults to 0.5.
numeric maximum probability
of the features persisting if projects are funded and successful.
Defaults to 0.9.
numeric minimum
probability of the features persisting if only the baseline project
is funded. Defaults to 0.01.
numeric maximum
probability of the features persisting if only the baseline project is
funded. Defaults to 0.4.
numeric of actions that are locked
into the solution. Defaults to 0.
numeric of actions that are locked
into the solution. Defaults to 0.
A list object containing the elements:
"projects"A tibble::tibble() containing
the data for the conservation projects. It contains the following
columns:
"name"character name for each project.
"success"numeric probability of each project
succeeding if it is funded.
"F1" ... "FN"numeric columns for each
feature, ranging from "F1" to "FN" where N
is the number of features, indicating the enhanced probability that
each feature will persist if it funded. Missing values (NA)
indicate that a feature does not benefit from a project being
funded.
"F1_action" ... "FN_action"logical
columns for each action, ranging from "F1_action" to
"FN_action" where N is
the number of actions (equal to the number of features in this
simulated data), indicating if an action is associated with a
project (TRUE) or not (FALSE).
"baseline_action"logical
column indicating if a project is associated with the baseline
action (TRUE) or not (FALSE). This action is only
associated with the baseline project.
"actions"A tibble::tibble() containing
the data for the conservation actions. It contains the following
columns:
"name"character name for each action.
"cost"numeric cost for each action.
"locked_in"logical indicating if certain
actions should be locked into the solution.
"locked_out"logical indicating if certain
actions should be locked out of the solution.
"features"A tibble::tibble() containing
the data for the conservation features (e.g. species). It contains the
following columns:
"name"character name for each feature.
"weight"numeric weight for each feature.
For each feature, this is calculated as the amount of time that
elapsed between the present and the features' last common ancestor.
In other words, the weights are calculated as the unique amount
of evolutionary history that each feature has experienced.
ape::phylo() phylogenetic tree for the
features.
The simulated data set will contain one conservation project for each features, and also a "baseline" (do nothing) project to reflect features' persistence when their conservation project is not funded. Each conservation project is associated with a single action, and no conservation projects share any actions. Specifically, the data are simulated as follows:
A specified number of conservation projects, features, and management actions are created.
Cost data for each action are simulated using a normal
distribution and the cost_mean and cost_sd arguments.
A set proportion of the actions are randomly set to be locked
in and out of the solutions using the locked_in_proportion and
locked_out_proportion arguments.
The probability of each project succeeding if its action is funded
is simulated by drawing probabilities from a uniform distribution with
the upper and lower bounds set as the success_min_probability
and success_max_probability arguments.
The probability of each feature persisting if various projects
are funded and is successful is simulated by drawing probabilities
from a uniform distribution with the upper and lower bounds set as the
funded_min_persistence_probability and
funded_max_persistence_probability arguments. To prevent
An additional project is created which represents the "baseline"
(do nothing) scenario. The probability of each feature persisting
when managed under this project is simulated by drawing probabilities
from a uniform distribution with the upper and lower bounds
set as the baseline_min_persistence_probability
and baseline_max_persistence_probability arguments.
A phylogenetic tree is simulated for the features using
ape::rcoal().
Feature data are created from the phylogenetic tree. The weights are calculated as the amount of evolutionary history that has elapsed between each feature and its last common ancestor.
Carwardine J, Martin TG, Firn J, Ponce-Reyes P, Nicol S, Reeson A, Grantham HS, Stratford D, Kehoe L, Chades I (2019) Priority Threat Management for biodiversity conservation: A handbook. Journal of Applied Ecology, 56: 481--490.
# create a simulated data set
s <- simulate_ptm_data(number_projects = 6,
number_actions = 8,
number_features = 5,
cost_mean = 100,
cost_sd = 5,
success_min_probability = 0.7,
success_max_probability = 0.99,
funded_min_persistence_probability = 0.5,
funded_max_persistence_probability = 0.9,
baseline_min_persistence_probability = 0.01,
baseline_max_persistence_probability = 0.4,
locked_in_proportion = 0.01,
locked_out_proportion = 0.01)
# print data set
print(s)
#> $projects
#> # A tibble: 7 × 16
#> name success F1 F2 F3 F4 F5 actio…¹ actio…² actio…³
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl> <lgl>
#> 1 project_1 0.931 NA 0.533 0.606 NA 0.878 TRUE TRUE TRUE
#> 2 project_2 0.899 0.691 NA 0.668 0.509 0.889 TRUE TRUE TRUE
#> 3 project_3 0.867 0.644 0.544 0.845 NA 0.547 TRUE TRUE TRUE
#> 4 project_4 0.892 NA 0.791 NA NA 0.666 TRUE FALSE TRUE
#> 5 project_5 0.876 0.556 NA 0.755 0.837 0.800 TRUE FALSE FALSE
#> 6 project_6 0.932 0.517 0.894 NA 0.520 0.663 TRUE FALSE TRUE
#> 7 baseline_… 1 0.0868 0.0692 0.321 0.376 0.303 FALSE FALSE FALSE
#> # … with 6 more variables: action_4 <lgl>, action_5 <lgl>, action_6 <lgl>,
#> # action_7 <lgl>, action_8 <lgl>, baseline_action <lgl>, and abbreviated
#> # variable names ¹action_1, ²action_2, ³action_3
#>
#> $actions
#> # A tibble: 9 × 4
#> name cost locked_in locked_out
#> <chr> <dbl> <lgl> <lgl>
#> 1 action_1 101. FALSE FALSE
#> 2 action_2 96.4 FALSE FALSE
#> 3 action_3 96.3 FALSE FALSE
#> 4 action_4 99.6 FALSE FALSE
#> 5 action_5 95.7 TRUE FALSE
#> 6 action_6 93.0 FALSE FALSE
#> 7 action_7 109. FALSE FALSE
#> 8 action_8 90.4 FALSE TRUE
#> 9 baseline_action 0 FALSE FALSE
#>
#> $features
#> # A tibble: 5 × 2
#> name weight
#> <chr> <dbl>
#> 1 F1 0.125
#> 2 F2 0.125
#> 3 F3 0.376
#> 4 F4 0.724
#> 5 F5 3.13
#>
#> $tree
#>
#> Phylogenetic tree with 5 tips and 4 internal nodes.
#>
#> Tip labels:
#> F1, F2, F3, F4, F5
#>
#> Rooted; includes branch lengths.
#>