Calculate periodicity metrics

This function takes in data for a recorder and calculates the periodicity metrics. All metrics are calculated within years.

periodicity(recorder_name, data, date_col = "date_start",
  recorder_col = "recorders", day_limit = 5)

Arguments

recorder_name	the name of the recorder for whom you want to calculate the metrics
data	the data.frame of recording information
date_col	the name of the column that contains the date. This must be formatted as a date
recorder_col	the name of the column that contains the recorder names
day_limit	the threshold number of days a recorder must be active before these metrics are estimated. If the number of active days for the recorder is less than this number then the function will return NA values.

Value

A data.frame with seven columns

• recorder - The name of the recorder, as given in the recorder_name argument

• periodicity - The median number of days elapsed between each pair of sequential active days. This describes the regularity with which people record.

• periodicity_variation - The standard deviation of the times elapsed between each pair of sequential active days

• median_streak - The median length of streaks, including streak lengths of 1

• sd_streak - The standard deviation of streaks lengths, including streak lengths of 1

• max_streak - the length of this recorders longest streak

• n_days - The number of dates on which this recorder made observations

Details

In this function a streak is defined as a series of consecutive days on which a person made observations. A streak of 1 is a single days recording in isolation. 2 is 2 days back-to-back, 3 is 3 days in a row and so on.

Examples

# NOT RUN {
# load example data
head(cit_sci_data)

# run for one recorder
P <- periodicity(recorder_name = 3007,
                data = cit_sci_data,
                date_col = 'date',
                recorder_col = 'recorder',
                day_limit = 5)

# Run the metric for all recorders
P_all <- lapply(unique(cit_sci_data$recorder),
               FUN = periodicity,
               data = cit_sci_data,
               date_col = 'date',
               recorder_col = 'recorder',
               day_limit = 5)

# summarise as one table
P_all_sum <- do.call(rbind, P_all)

hist(P_all_sum$max_streak)
# }