Process hurricane data

process_data.R

# Process hurricanes data
hurricanes <-
  hurricanes %>%
  mutate(year = Season) %>%
  # Keep only those since 1917
  filter(year >= 1917) %>%
  mutate(date_time = ISO_time,
         lon = Longitude,
         lat = Latitude,
         id = Serial_Num,
         wind = `Wind(WMO)`,
         pressure = `Pres(WMO) Percentile`) %>%
  arrange(date_time) %>%
  # Keep out those which cross the longitudinal start-over
  group_by(id) %>%
  mutate(cross_over = any(lon < -100) & any(lon > 100)) %>%
  ungroup %>%
  filter(!cross_over) %>%
  # Keep only those in the Atlantic basin
  group_by(id) %>%
  filter(any(lon >-90) & any(lon < -10),
         any(lat > 10) & (any(lat < 40))) %>%
  # Get max wind
  mutate(max_wind = max(wind)) %>%
  ungroup %>%
  # Make ids simple
  mutate(id = as.numeric(factor(id)))

# Get an index
hurricanes$index <- 1:nrow(hurricanes)
hurricanes <- hurricanes %>%
  group_by(id) %>%
  mutate(max_index = max(index)) %>%
  ungroup

# Define ids of storms
ids <- sort(unique(hurricanes$id))
years <- sort(unique(hurricanes$year))
max_indices <- sort(unique(hurricanes$max_index))