Storm data preparation
Since I wnat to focus on the analysis of the storm data, I use R’s hurdat package, which does a lot of the cleaning for those who want to save time (vignette).
Reading the data
I read in two data sets: one for hurricanes (Atlantic ocean) and one for typhoons (northeast and central Pacific ocean). The original data format is the revised Atlantic hurricane database (HURDAT2, see here for details). You probably want to have a look at the decription that is provided here. To each data set I add a variable that indicates the respective ocean it belongs to.
# Data for Atlantic storms.
dd.atlantic <- get_hurdat(basin = "AL") %>%
mutate(Ocean = "Atlantic")
# Data for northeast and north-central Pacific storms.
dd.pacific <- get_hurdat(basin = "EP") %>%
mutate(Ocean = "Pacific")
I combine both data sets to a single one.
dd.org <- bind_rows(dd.atlantic, dd.pacific)
str(dd.org)
## Classes 'tbl_df', 'tbl' and 'data.frame': 76559 obs. of 22 variables:
## $ Key : chr "AL011851" "AL011851" "AL011851" "AL011851" ...
## $ Name : chr "UNNAMED" "UNNAMED" "UNNAMED" "UNNAMED" ...
## $ DateTime: POSIXct, format: "1851-06-25 00:00:00" "1851-06-25 06:00:00" ...
## $ Record : chr NA NA NA NA ...
## $ Status : chr "HU" "HU" "HU" "HU" ...
## $ Lat : num 28 28 28 28.1 28.2 28.2 28.3 28.4 28.6 29 ...
## $ Lon : num -94.8 -95.4 -96 -96.5 -96.8 -97 -97.6 -98.3 -98.9 -99.4 ...
## $ Wind : int 80 80 80 80 80 70 60 60 50 50 ...
## $ Pressure: int NA NA NA NA NA NA NA NA NA NA ...
## $ NE34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NE50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NE64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ Ocean : chr "Atlantic" "Atlantic" "Atlantic" "Atlantic" ...
Next, I make a couple of adjustments so that the data set matches my needs for analysis:
- Year: to summarise/describe data by year.
- DateTimeSameYear: when you only want to explore seasonal occurances.
- Duration: giving amount of time s storm lasted.
- WindKPH: wind speed in km per hour.
- KeyPlus: a combination of Key and Name.
- SaffirSimpson: assigning storms a value on the Saffir-Simpson scale.
dd <- dd.org %>%
mutate(Year = year(DateTime),
DateTimeSameYear = as.POSIXct(strptime(format(DateTime, "%m/%d %H:%M:%S"), "%m/%d %H:%M:%S")),
WindKPH = round(Wind * 1.609344, 0),
Duration = NA, # Needed at a latter point.
KeyPlus = paste(Name, " (",Key,")", sep = ""),
SaffirSimpson = ifelse(WindKPH %in% 0:62, "Tropical depression",
ifelse(WindKPH %in% 63:118, "Tropical storm",
ifelse(WindKPH %in% 119:153, "1",
ifelse(WindKPH %in% 154:177, "2",
ifelse(WindKPH %in% 178:208, "3",
ifelse(WindKPH %in% 209:251, "4", "5")))))))
There is one manipulation that is critical for plotting (at least) with Leaflet for R: All longitude values have to be positive. Otherwise Leaflet gets confused and the data “jumps” by 360\degree.
dd$Lon <- ifelse(dd$Lon < 0, dd$Lon + 360, dd$Lon)
Then, for some variables it makes more sense if they are declared as factors.
dd[c("Key",
"Year",
"KeyPlus",
"Status",
"Ocean",
"SaffirSimpson")] = lapply(dd[c("Key",
"Year",
"KeyPlus",
"Status",
"Ocean",
"SaffirSimpson")], as.factor)
Finally, I calculate the duration of each storm.
# ids <- as.character(unique(dd$ID))
for (i in as.character(unique(dd$Key))) {
dd.i <- filter(dd, Key == i)
for (n in 1:length(dd.i$Key)) {
# '%--%' is the interval operator returning the time difference (lubridate).
dd.i$Duration[n] <- round(as.numeric(as.duration(dd.i$DateTime[1] %--% dd.i$DateTime[n]))/60/60/24, 2)
}
dd$Duration[which(dd$Key %in% i)] <- dd.i$Duration
}
There you go!
## Classes 'tbl_df', 'tbl' and 'data.frame': 76559 obs. of 28 variables:
## $ Key : Factor w/ 2900 levels "AL011851","AL011852",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ Name : chr "UNNAMED" "UNNAMED" "UNNAMED" "UNNAMED" ...
## $ DateTime : POSIXct, format: "1851-06-25 00:00:00" "1851-06-25 06:00:00" ...
## $ Record : chr NA NA NA NA ...
## $ Status : Factor w/ 12 levels "DB","ET","EX",..: 4 4 4 4 4 4 11 11 11 11 ...
## $ Lat : num 28 28 28 28.1 28.2 28.2 28.3 28.4 28.6 29 ...
## $ Lon : num 265 265 264 264 263 ...
## $ Wind : int 80 80 80 80 80 70 60 60 50 50 ...
## $ Pressure : int NA NA NA NA NA NA NA NA NA NA ...
## $ NE34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW34 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NE50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW50 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NE64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SE64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ SW64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ NW64 : int NA NA NA NA NA NA NA NA NA NA ...
## $ Ocean : Factor w/ 2 levels "Atlantic","Pacific": 1 1 1 1 1 1 1 1 1 1 ...
## $ Year : Factor w/ 166 levels "1851","1852",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ DateTimeSameYear: POSIXct, format: "2018-06-25 00:00:00" "2018-06-25 06:00:00" ...
## $ WindKPH : num 129 129 129 129 129 113 97 97 80 80 ...
## $ Duration : num 0 0.25 0.5 0.75 0.88 1 1.25 1.5 1.75 2 ...
## $ KeyPlus : Factor w/ 2900 levels "ABBY (AL011968)",..: 1598 1598 1598 1598 1598 1598 1598 1598 1598 1598 ...
## $ SaffirSimpson : Factor w/ 7 levels "1","2","3","4",..: 1 1 1 1 1 7 7 7 7 7 ...
Storing the data set
I save the workspace for later use and exploration/analysis.
save.image("StormDataWorkSpace.RData")