In this earlier post we analysed the location of meteorite impacts from this dataset, including plotting their fall locations on a globe. Forming part of the analysis was this interactive globe visualization below, which plots the location and age of meteorites.

In this post I am going to show how to create this globe with code. I then go on to describe other options and variations on the same theme.

Preparing and plotting the data

The code below does everything we need. Stepping through each section:

• First, import the data then convert from the year the meteorite fell to an age in years.
• Tidy the data frame to collect just the variables that we need.
• Convert age into a number from 1 to 10 (the lowest 10% of ages map to 1, the next 10% to 2 .. etc), then map those numbers to shades of red.
• Finally, plot the data on the globe where val (the mass) determines the length of each line and fixed pointsize determines the line thickness.

library(threejs)
library(flipChartBasics)

# Read the data and calculate age in years
x = read.csv("https://data.nasa.gov/api/views/gh4g-9sfh/rows.csv")
current = as.numeric(format(Sys.Date(), "%Y"))
x$age = current - as.numeric(substr(x$year, 7, 10))

# Filter the required information
x = x[ , c("reclong", "reclat", "mass..g.", "age")]
colnames(x) = c("long","lat","value", "age")

# Set colors on a scale of 1 to 10 by percentile
colors = as.numeric(cut(x$age,
breaks = quantile(x$age, probs = seq(0, 1, 0.1), 
                  include.lowest = TRUE, na.rm = TRUE)))
palette = ChartColors(10, "Reds", reverse = TRUE)
colors = palette[colors]

# Plot the data on the globe
globejs(lat = x$lat,
        long = x$long,
        val = 2 * log(x$value),
        color = colors,
        pointsize = 0.5,
        atmosphere = TRUE)

The output is slightly different from the original in the meteorites post because we have not excluded any data. You can spin the globe around with your mouse and zoom in or out.

Capital cities

As a second example, we’ll plot capital city data available here. The data is imported with the DownloadXLSX function from the flipAPI package.

The Population (thousands) column is imported as a factor. We amend it to be a number. The code to make the globe is broadly similar to that above.

library(threejs)
library(flipChartBasics)
library(flipAPI)

# Make a data.frame of the required information
url <- "https://esa.un.org/unpd/wup/cd-rom/WUP2014_XLS_CD_FILES/WUP2014-F13-Capital_Cities.xls"
x = DownloadXLSX(url, skip = 15)
x = x[, c("Longitude", "Latitude", "Population (thousands)", "Capital City")]
names(x) = c("long","lat", "population", "city")

# Convert population to numeric
x$population = as.character(x$population)
x$population[x$population == "\u2026"] = 0 # remove ellipsis
x$population = as.numeric(x$population)

# Set colors according to first letter of the city name
first.letters = sapply(substring(x$city, 1, 1), 
                       utf8ToInt) - utf8ToInt("A") + 1
palette = ChartColors(26, "Blues")
colors = palette[first.letters]

# Plot the data on the globe
earth = "http://eoimages.gsfc.nasa.gov/images/imagerecords/73000/73909/world.topo.bathy.200412.3x5400x2700.jpg"
globejs(img = earth,
        lat = x$lat,
        long = x$long,
        val = 10 * log(x$population), 
        color = colors,
        pointsize = 5,
        atmosphere = FALSE,
        bg = "white")

A few notable differences are:

• Addition of a more colorful background world image via the img argument.
• Line lengths related to population and colors according to the first letter of the city name (light blue = A, dark = Z).
• Removal of atmosphere, increase of pointsize and amending the default background (bg) color.

Finally, the globe image can be changed. If you wanted to confuse people by presenting the same information with an arbitrary citrus theme, you could make the following, again using the img argument.

The globes are created with the threejs package. You can see and modify the examples in this post by following this link.