Prepping my data

I had been keeping a list in Notes of the places I have been to. To prep for this tasks, I copied them into Google Sheets and quickly assigned them one of three categorical values: visit, transit, long_term. I then saved this as a .csv file in my working directory.

1. Load the countries.csv file and store it as an object data
2. Prep for the next step by eplacing the country names with their ISO3 codes using the countrycode package

data <- read.csv("countries.csv", header = TRUE)
  
data$country_name <- countrycode(data$country_name, origin = 'country.name', destination = 'iso3c')

## Warning in countrycode_convert(sourcevar = sourcevar, origin = origin, destination = dest, : Some values were not matched unambiguously: England, Scotland

Combine Scotland and England into one entry

The following pipe will fix my putting Scotland and England in my list as separate countries (“Doh!”). Here’s a play-by-play of what it does:

1. Remove the subset of rows that return NA (in this case Scotland and England0
2. Add a new row for the United Kingdom
3. Replace the United Kingdom value with its ISO3 code
4. Store all as an updated data frame named newdata

newdata <- data[-which(is.na(data)), ] %>%
  add_row(country_name = countrycode("United Kingdom", origin = 'country.name', destination = 'iso3c'), trip_type = "visit")

1) Summary of trip_type

While I knew there were only three categories, it’s always nice to double-check in case I made any typos that might mess up my maps. The easiest way to do this is to take a look at the summary stats of the variable in question.

table(newdata$trip_type)

## 
## long_term   transit     visit 
##         8         7        27

length(newdata$trip_type)

## [1] 42

2) Option 1: Create Map Using rworldmap

Using the rworldmap package, I will use the joinCountryData2map() function to join my nice and clean data frame of countries I’ve been to with the country map data in the rworldmap package.

country_map <- joinCountryData2Map(newdata, 
                                   joinCode = "ISO3",
                                   nameJoinColumn = "country_name")

## 42 codes from your data successfully matched countries in the map
## 0 codes from your data failed to match with a country code in the map
## 201 codes from the map weren't represented in your data

Now that all the info is combined into the country_map object, I will use the mapCountryData() function to start mapping. I’ll set the colors to match the ones used throughout my website, and assign countries I have yet to step foot into an opaque gray.

Note: the order of colors in the colourPalette argument match up to the categories of the trip_type variable, since R arranges character elements in alphabetical order by default (lived, transit, visit).

map1 <- mapCountryData(country_map, nameColumnToPlot="trip_type", 
                       catMethod = "categorical", 
                       missingCountryCol = gray(.8),
                       colourPalette = c("black", "turquoise", "blue"), 
                       mapTitle = "Places I've been to", borderCol = "white")

Meh. It works, but it’s not the aesthetically-pleasing result I was hoping for.

I could add the argument addLegend = FALSE to remove the clunky legend, then use the addMapLegend function to create a custom one. Beyond that, however, there are few styling options in the rworldmap package.

To really get something pretty and presentable, I will need to use some more robust mapping and data visualization packages: ggplot2 and ggirafe.

3) Option 2: Create map using ggplot2

Create a base map by retrieving spacial information about, well, every country on Earth using the rnaturalearth and rnaturalearthdata packages. I will also need to install the sf package, which will help with encoding and working with the spacial data retrieved using the first packages.

suppressPackageStartupMessages(library(rnaturalearth))
suppressPackageStartupMessages(library(rnaturalearthdata))
suppressPackageStartupMessages(library(sf))

## Warning: package 'sf' was built under R version 4.1.2

First, I will create an sf object, or spacial data frame, using the ne_countries() function in the rnaturalearth package.

Once that info is stored into a data frame called world, I can add a new variable or column to add information about whether I visited, lived in, or transited through each of the countries using the merge() function. There’s just one problem: I haven’t changed the variable name of country_names yet, so R won’t be able to find a common variable to merge with world.

I could change all the ISO3 values back to country names and rename the column to match any one of the sovereignt column. Instead, I decided to change it to iso_a3 instead, which was more accurate.

world <- ne_countries(scale = "medium", returnclass = "sf") 

newdata2 <- rename(newdata, iso_a3 = country_name)

world2 <- merge(world, newdata2, all = TRUE)

Now that the object world2 is ready and complete with all the countries I’ve been to, we can get to mapping - again. While there’s plenty you can do with ggplot2, I wanted some extra styles. I installed and loaded the ggthemes package below, which has some great templates you can tweak. I chose the theme_wsj, inspired by the styles of the Wall Street Journal.

suppressPackageStartupMessages(library(ggthemes))
suppressPackageStartupMessages(library(ggplot2))


WorldMap2 <- ggplot(data = world2) +
    geom_sf(aes(fill = trip_type)) +
    theme_wsj(color = "white", base_family = "mono") + 
    ggtitle("Places I've been to", subtitle = paste0("(", sum(!is.na(world2$trip_type)), " countries)")) +
    scale_fill_manual(breaks = c("long_term", "visit", "transit"), 
                      values=c("#5c0099", "#fdc500", "#DDDDDD"),
                      labels=c("Lived in", "Visited", "Traveled through")) +
    guides(fill=guide_legend(title=NULL))

WorldMap2

Much better. As pretty as this is, I want it to be more dynamic. Not just because it looks cool, but there are some practical benefits like:

1. People would be able to zoom and see smaller European countries
2. For the less geographically gifted (no judgment, I myself put Scotland and England as separate countries), I want it to be easy to identify which country is which