A Package Full o’ Pirates & Makin’ Interactive Pirate Maps in arrrrrRstats
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
Avast, me hearties! It’s time four t’ annual International Talk Like a Pirate Day #rstats post!
(OK, I won’t make you suffer continuous pirate-speak for the entire post)
I tried to be a bit more practical this year and have two treasuRe chests for you to (hopefully) enjoy.
A Package Full o’ Pirates
I’ve covered the Anti-shipping Activity Messages (ASAM) Database before for TLAPD before but getting, updating and working with the data has more pain points than it should, so I wrapped a small package around it.
Here’s how to get all pirate attacks this year (2015) so far:
# devtools::install_github("hrbrmstr/asam")
library(asam)
data(asam_shp)
pirates <- subset(asam_shp,
grepl("pirate", Aggressor, ignore.case=TRUE) &
format(DateOfOcc, "%Y") == "2015")
nrow(pirates)
## [1] 78 |
It looks like there have been 78 registered pirate attacks this year. The National Geospatial Intelligence Agency (NGIA) marks the attacks by lat/lon and also by region/subregion, and I managed to obtain the official polygons for these regions, so we can plot these attacks on a world map and also show the subregions:
library(ggplot2)
# get the ASAM subregion polygons
subregions <- asam_subregions()
subregions_map <- fortify(subregions)
# get the world map
world <- map_data("world")
# get the points for the pirate attack occurrences
pirate_pts <- data.frame(pirates)
gg <- ggplot()
# world map layer
gg <- gg + geom_map(data=world, map=world,
aes(x=long, y=lat, map_id=region),
color="black", fill="#e7e7e7", size=0.15)
# ASAM regions layer
gg <- gg + geom_map(data=subregions_map, map=subregions_map,
aes(x=long, y=lat, map_id=id),
color="white", fill="white", size=0.15, alpha=0)
# attacks
gg <- gg + geom_point(data=pirate_pts, color="black", fill="yellow",
aes(x=coords.x1, y=coords.x2), shape=21)
gg <- gg + xlim(-170, 170)
gg <- gg + ylim(-58, 75)
gg <- gg + coord_map("mollweide")
gg <- gg + theme_map()
gg <- gg + theme(panel.background=element_rect(fill="steelblue"))
gg |
There is quite a bit more data than just location, though and we can work with it much better in an interactive map.
Makin’ Interactive Pirate Maps
Now, what makes the following an interactive pirate map is not so much the fact that we’ll be plotting points of pirate attacks on a Leaflet map, but we’ll also be using a pirate treasure map theme on the Leaflet map.
Let’s start with showing how to use a general pirate map theme before adding in the ASAM data.
You’ll have to pause here and head on over to MapBox to register for a (free) account. You’ll need to go through the gyrations to eventually get a public token and mapbox id to use the pirate map tiles they have. I store those in my .Renviron so I don’t have to cut/paste inane strings when I need to use this, or other, APIs or need to keep them from prying eyes. Since MapBox exposes these strings in GET call URLs, the use of environment variables is strictly for convenience in this case.
library(leaflet)
mapbox_public_token <- Sys.getenv("MAPBOX_PUBLIC_TOKEN")
mapbox_map_id <- Sys.getenv("PIRATE_MAP_ID")
mapbox_url <- "https://a.tiles.mapbox.com/v4/%s/{z}/{x}/{y}.png?access_token=%s"
mapbox_tiles_template <- sprintf(mapbox_url, mapbox_map_id, mapbox_public_token) |
Now, what good is a pirate map without an ‘X’ marking the spot for some treasure. For that we’ll need an ‘X’:
in a format we can use with Leaflet:
x_marker <- icons("http://rud.is/dl/x.png",
iconHeight=64, iconWidth=64,
iconAnchorX=32, iconAnchorY=32) |
Now, we can display a pirate map for all scurvy dogs to see:
leaflet() %>% addTiles(mapbox_tiles_template) %>% setView(lng=-50.9249, lat=45.68929, zoom=3) %>% addMarkers(-70.2667, 43.6667, icon=x_marker) |
NOTE: I have not buried treasure in Portland, Maine, but go nuts digging at that location if you still want to.
Pirates on Pirate Maps
We can make a [crude] interactive ASAM browser by combining our data from above with our new, pirate-y mapping capabilities:
library(asam)
library(sp)
library(dplyr)
library(leaflet)
data(asam_shp)
dat <- subset(asam_shp,
DateOfOcc > as.Date("2015-01-01") &
grepl("pirate", Aggressor, ignore.case=TRUE))
# could also do data.frame(dat)
dat <- bind_cols(dat@data, data.frame(coordinates(dat), stringsAsFactors=FALSE)) |
We’ll build a popup with the ASAM incident description fields and add it and the pirate incident points to a pirate-themed Leaflet map:
popup_template <- '<div style="background:#f3e0b5; padding:10px"><b>Date:</b> %s
<span style="float:right"><a target="_blank"
href="http://msi.nga.mil/NGAPortal/msi/query_results.jsp?MSI_queryType=ASAM&MSI_generalFilterType=SpecificNumber&MSI_generalFilterValue=%s&MSI_additionalFilterType1=None&MSI_additionalFilterType2=-999&MSI_additionalFilterValue1=-999&MSI_additionalFilterValue2=-999&MSI_outputOptionType1=SortBy&MSI_outputOptionType2=-999&MSI_outputOptionValue1=Date_DESC&MSI_outputOptionValue2=-999&MSI_MAP=-999">ASAM Record</a>
</span><br/>
<b>Victim:</b> %s<br/>
<b>Description:</b> %s</div>'
nona <- function(x) ifelse(is.na(x), " ", x)
pirate_pops <- sprintf(popup_template,
dat$date,
gsub("-", "_", dat$Reference),
dat$Victim,
paste0(nona(dat$Descript),
nona(dat$Desc1), nona(dat$Desc2), nona(dat$Desc3),
nona(dat$Desc4), nona(dat$Desc5), nona(dat$Desc6),
sep=" "))
mapbox_public_token <- Sys.getenv("MAPBOX_PUBLIC_TOKEN")
mapbox_map_id <- Sys.getenv("PIRATE_MAP_ID")
mapbox_url <- "https://a.tiles.mapbox.com/v4/%s/{z}/{x}/{y}.png?access_token=%s"
mapbox_tiles_template <- sprintf(mapbox_url, mapbox_map_id, mapbox_public_token)
leaflet() %>%
addTiles(mapbox_tiles_template) %>%
setView(lng=-50.9249, lat=45.68929, zoom=3) %>%
addCircles(dat$coords.x1, dat$coords.x2, radius=300,
color="#664c1f", popup=pirate_pops) |
Select any of the circle marks and you’ll get a popup with a description and link to the official ASAM record (like this):
Fin
I’m not sure when I’ll get back to the asam package, but it could use some attention. The Aggressor field could be auto-cleaned to make it more usable and a dplyr-esque interface could be developed to select incidents. Also, since it includes a shapefile of subregions, that could also be used to do more spatial-oriented analyses of the incidents. It’s all there for any pirate lackey to pilfer.
Drop a note in the comments if you have any of your own pirate-y creations or an issue on github for feature requests & bug reports.
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.