old_data.R

### Legacy data from Jasper National Park

old_data <- read_csv("./assets/jasper_legacy_data.csv")

### Three different versions of legacy data tiedied and formatted. One is max count, one is mean across the ten minute period and one is mean across the 3 minute period. Goal is to determine the differences between the uses cases for modelling.

# Function to preprocess data
old_s_max_all <- old_data |>
    dplyr::select(observer, pointID, latitude, longitude, elevation, ecoregion, temperature, sky, wind,
                  survey_duration, date, species_code, species_common_name, TTFD, start_time,
                  abundance3.3:abundance10, max.species.individuals) |>
    distinct() |>
    rename(location = pointID) |>
    mutate(location = str_to_upper(location),
           location = gsub("([A-Z]+)([0-9]+)", "\\1-\\2", location),
           location = case_when(
             grepl("VALLEY", location) ~ gsub("(\\w+)-([0-9])", "\\1\\2-", location),
             TRUE ~ location
           )) |>
    mutate(recording_date_time = lubridate::mdy(date) + seconds(as.numeric(hms(start_time)))) |>
    relocate(c(location, latitude, longitude, recording_date_time, observer, species_common_name)) |>
    group_by(location, latitude, longitude, recording_date_time, elevation, ecoregion, temperature, sky, wind,
             survey_duration, observer, species_code, species_common_name, TTFD) |>
    summarise(individual_order = as.numeric(max(across(starts_with("abundance"))))) |>
    ungroup() |>
    dplyr::select(location, latitude, longitude, ecoregion, elevation, temperature, sky, wind, recording_date_time, observer, species_code,
                  species_common_name, TTFD, individual_order) |>
    distinct() |>
    rename(detection_time = TTFD) |>
    mutate(detection_time = floor(detection_time) * 60 + (detection_time %% 1) * 100) |>
    mutate(location = case_when(
      grepl("^POBOK", location) ~ gsub("POBOKTON", "POBOKTAN", location),
      grepl("^ELY", location) ~ gsub("ELYISUM", "ELYSIUM", location),
      TRUE ~ location
    )) |>
    mutate(organization = "JNP", .before = location, data_type = "legacy") |>
    mutate(year = year(recording_date_time))

old_s_max_temps <- old_s_max_all |>
  filter(year %in% c(2021:2023)) |>
  dplyr::select(location, year, temperature, sky, wind) |>
  rename(temp = temperature,
         sk = sky,
         wi = wind) |>
  distinct()

old_s_max <- old_s_max_all |>
    filter(!year %in% c(2021:2023)) |>
    dplyr::select(-year)