Species occurrence data to CartoDB

We have previously written about creating interactive maps on the web from R, with the interactive maps on Github. See here, here, here, and here.

A different approach is to use CartoDB, a freemium service with sql interface to your data tables that provides a map to visualize data in those tables. They released an R interace to their sql API on Github here - which we can use to make an interactive map from R.

We’ll first get some data from GBIF, ~500 occurrences of Puma concolor in the US, then push that data to a CartoDB table. There are a couple more non-programmatic steps in this workflow than with pushing geojson file to Github as outlined in the previous linked above (i.e., going to the CartoDB site and making a visualization, and making it public).

Install packages

install.packages("devtools")
library(devtools)
install_github("rgbif", "ropensci", ref = "newapi")
install_github("cartodb-r", "Vizzuality", subdir = "CartoDB")

Load em

library(rgbif)
library(CartoDB)

Get some data from GBIF

Here, we’ll get data for Puma concolor, the hello, world for biodiversity data.

key <- gbif_lookup(name = "Puma concolor", kingdom = "animals")$speciesKey
data <- occ_search(taxonKey = key, limit = 500, georeferenced = TRUE, country = "US",
    return = "data")
head(data)

##                             name longitude latitude
## 1 Puma concolor (Linnaeus, 1771)    -108.9    32.70
## 2 Puma concolor (Linnaeus, 1771)    -108.0    32.88
## 3 Puma concolor (Linnaeus, 1771)    -105.5    32.95
## 4 Puma concolor (Linnaeus, 1771)    -107.8    33.61
## 5 Puma concolor (Linnaeus, 1771)    -107.5    33.00
## 6 Puma concolor (Linnaeus, 1771)    -106.5    36.69

str(data)

## 'data.frame':	500 obs. of  3 variables:
##  $ name     : Factor w/ 7 levels "Animalia","Carnivora",..: 7 7 7 7 7 7 7 7 7 7 ...
##  $ longitude: num  -109 -108 -105 -108 -107 ...
##  $ latitude : num  32.7 32.9 33 33.6 33 ...

Great, we have some data. Now let’s make a map.

Push data up to CartoDB

I frist crated a table in my CartoDB account named pumamap. Then, I need to initialize the connection with CartoDB with my account name and API key. Note that I am pulling up my key from my .Rprofile file on my machine for ease and so it’s not revealed to you :)

key = getOption("mycartodbkey")
cartodb("recology", api.key = key)

Now we need to push data to our pumamap table using the function cartodb.row.insert. It accepts one row of data, so we’ll pass each row of data with an lapply call.

rows <- apply(data, 1, as.list)
lapply(rows, function(x) cartodb.row.insert(name = "pumamap", columns = list("name",
    "longitude", "latitude"), values = x))

After the upload is finished, I had to make sure the table was georeferenced, and played with settings to suit my style. And then I made a visualization from the pumamap dataset and made it public. And that’s it! You can find the map, and it can be embedded:

And we can examine a row from the table in our CartoDB account with a single line of code

cartodb.row.get(name = "pumamap", cartodb_id = 10)

##   cartodb_id                           name description
## 1          1 Puma concolor (Linnaeus, 1771)        NULL
##                 created_at               updated_at
## 1 2013-11-03T06:40:12+0100 2013-11-03T06:46:55+0100
##                                             the_geom the_geom_webmercator
## 1 0101000020E610000089247A19C5365BC08C15359886594040                 NULL
##   latitude longitude
## 1     32.7    -108.9