Convert camera trap data into a recurrent event format

Takes a camera trap dataset as input (site ID, timestamp and species name) and return a recurrent event dataset. The function creates a survey after each observation of the primary species for a defined maximum duration. Within each survey, all the observations of the secondary species are converted into recurrent events. The returned dataframe can be directly used for recurrent event analysis.

ct_to_recurrent(
  data,
  primary,
  secondary,
  survey_duration = 10,
  datetime_var = "DateTime",
  species_var = "Species",
  site_var = "Site",
  tertiary = NULL,
  survey_end_date = NULL
)

Arguments

data: A dataframe containing the site, the timestamp and the species information
primary: character A vector of one or several species names considered to affect the secondary species
secondary: character(1) Name of the species affected by the primary (only one name allowed).
survey_duration: integer(1) Maximum duration of the survey (in days, e.g. "7")
datetime_var: Name of the variable that contains date time information.
species_var: character(1) Name of the variable that contains species information.
site_var: Name of the variable that contains Site ID.
tertiary: character Names of the species in species_var that should not be considere primary or secondary.
survey_end_date: Date Date of study end (e.g. "01-01-2001")

Value

A dataframe containing:

Site: Camera-trap site ID
ID: Survey ID
primary: primary species starting the survey
datetime_primary: Timestamp of the start of the survey (i.e. of the primary species observation)
DateTime: Timestamp of the recurrent event (i.e. of the secondary species observation)
t.start: Time since the start of the survey
t.stop: Time since the last event, which can be the primary species or a previous secondary species
event: Binary, 1 = recurrent event, 0 = censoring event
enum: Event number within the survey
Event_type: Either the recurrent event, giving the name of the secondary species, or a censoring event and the reason of censoring
datetime_primary: Timestamp of survey start (primary observation)
...: Any covariate attached to the dataset