Getting Started in ggplot

Josh Allen

Department of Political Science at Georgia State University

8/29/22

Where We Have Been

plot(penguins$bill_length_mm,
   penguins$body_mass_g,
   xlab = "Bill Length(mm)",
   ylab = "Body Mass(g)")
   abline(lm(body_mass_g~bill_length_mm, data = penguins))

We have seen how to make plots in base R they take this general form

Expanding What We Know

par(mfrow = c(2,2))
plot(penguins$bill_length_mm,
        penguins$body_mass_g,
    xlab = "Bill Length(mm)",
   ylab = "Body Mass(g)")
hist(penguins$bill_length_mm,
  xlim = c(30, 60))
plot(density(penguins$bill_length_mm))
plot(penguins$bill_length_mm,
    penguins$body_mass_g,
    xlab = "Bill Length(mm)",
    ylab = "Body Mass(g)")
abline(lm(body_mass_g~bill_length_mm, data = penguins))

you have to add something like jpeg(“name_of_plot.jpg”)

and then add dev.off() at the end

Research Data Services

Our Team

Also, we do not have the capacity to provide pre-scheduled frequent software tutoring that is divorced from a specific assignment or software troubleshooting task at hand – in other words, we are also not generalized software tutors.

Our one-on-one software assistance is available to help you troubleshoot specific tasks or assignments. If you are seeking generalized software help, we direct you to our live workshops and our recorded workshops to gain the foundational skills for using various analytical software. Then, when or if the time comes that you have targeted software questions or specific issues to tackle related to a course assignment or research project, please feel free to contact us for one-on-one support.

Get Ready Badges

https://research.library.gsu.edu/dataservices/data-ready

These are awesome to share on social media i.e. linkedin which is a good signal to potential employers that you know this stuff

How To Get the Badges

The Importance of Graphing

Why visualize your data?

mean(graph_dat$x)

[1] 54.2657

sd(graph_dat$x)

[1] 16.713

mean(graph_dat$y)

[1] 47.8351

sd(graph_dat$y)

[1] 26.84777

cor(graph_dat$x, graph_dat$y)

[1] -0.06601891

Lets take a quick look at the data that I have loaded in. They look pretty similar for the most part and aren’t correlated.

The Dino Strikes

The Datasaurus Dozen

Why ggplot2?

The transferable skills from ggplot2 are not the idiosyncrasies of plotting syntax, but a powerful way of thinking about visualization, as a way of mapping between variables and the visual properties of geometric objects that you can perceive.

– Hadley Wickham

Why ggplot2?

• You have probably heard of it but why use it?

• Once we understand the “grammar” making figures becomes a lot easier

• Tons of organizations use it

• Flexibility

  • Tons ways to customize appearance
  • Lots of functions
  • Lots of extensions

• Reproducibility

  • Doesn’t require you to remember each input from a drop down menu
  • Defaults to universally usable formats
  • Replaces itself automatically in your directory

ggplot2 is probably one of the most well known and event among many of the staunchest base r users well loved an well used libraries in the R ecosystem.

ggplot2 predates most of the tidyverse and the gg part refers to the grammar of graphics, more on that later. Much like writing a sentence there is a proper grammar to it. Once we understand the rules we can construct lots of different “sentences”.

I think an important theme so to say is that once we start to understand R it benefits your workflow because we can simply tweak a few things and then BAM it will just update. If you are a Latex user then bam the same plot just updates. If you are a word user then those benefits are limited but ggplots play nicely with Word. No having to put replace everywhere or screenshot things or working with propietary formats. If you need to you

The Grammar of Graphics

Grammar

“Good grammar is just the first step of creating a good sentence”

• How is the data related to the figure on the right?

Building the Plot

Body Weight of Penguins and Bill Length

• Penguins

• Species

• Island

I apologize for some weird plotting stuff. I am still working out some kinks with how the presentation is styled. I wanted to try to give the plots and the code the same size on the slides. However that resulted in some weird decisions from quarto

Building the Plot

Body Weight of Penguins and Bill Length

• Penguins

• Species

• Island

There is a grammar to what is going on

Building the Plot

Body Weight of Penguins and Bill Length

• Penguins

• Species

• Island

So How Did We go From?

This

To This

Making Plots

The Grammar

Component	Function	Explanation
Data	ggplot(data)	The raw data that you want to visualise.
Aesthetics	aes()	Aesthetic mappings between variables and visual properties.
Geometries	geom_*()	The geometric shapes representing the data.
Statistics	stat_*()	The statistical transformations applied to the data.
Scales	scale_*()	Maps between the data and the aesthetic dimensions.
Coordinate System	coord_*()	Maps data into the plane of the data rectangle.
Facets	facet_*()	The arrangement of the data into a grid of plots.
Visual Themes	theme() and theme_*()	The overall visual defaults of a plot.

This is what it looks like when we start to think of the plot. Each of these are columns in our dataset that we plop into aes. Our x goes in the first slot of aes and the second slot is our y. I typically specify this myself just cuz I like to do that but that is a personal decision. The rest you need to tell ggplot to do coordinates Combine the two position aesthetics to produce a 2d position on the plot. The position aesthetics are called x and y, but they might be better called position 1 and 2 because their meaning depends on the coordinate system used. For example, with the polar coordinate system they become angle and radius (or radius and angle), and with maps they become latitude and longitude. scales take your data and turn it into something that you can see, like size, colour, position or shape. They also provide the tools that let you interpret the plot: the axes and legends. You can generate plots with ggplot2 without knowing how scales work

Where do they go?

ggplot() +
  geom_point(data = penguins,
   aes(
  x = bill_length_mm, 
  y = body_mass_g,
  shape = species, 
  color = island),
  size = 3)

To actually make the plot you need to specify the columns you want to “map” onto the plot. Map is kind of just the way that it is said in ggplot world. This is just us saying we want these columns in our

Plotting Data

country	continent	year	lifeExp	pop	gdpPercap
Afghanistan	Asia	1952	28.801	8425333	779.4453
Afghanistan	Asia	1957	30.332	9240934	820.8530
Afghanistan	Asia	1962	31.997	10267083	853.1007
Afghanistan	Asia	1967	34.020	11537966	836.1971
Afghanistan	Asia	1972	36.088	13079460	739.9811
Afghanistan	Asia	1977	38.438	14880372	786.1134
Afghanistan	Asia	1982	39.854	12881816	978.0114
Afghanistan	Asia	1987	40.822	13867957	852.3959
Afghanistan	Asia	1992	41.674	16317921	649.3414
Afghanistan	Asia	1997	41.763	22227415	635.3414

Here is your shell script

## be sure you have done 
## install.packages("gapminder")
## library(gapminder)

ggplot() +
  geom_point(data = gapminder, mapping = aes(x = gdpPercap, y = lifeExp))

To help you get your hands dirty I will give y’all a shell script to work with. Press o on your keyboard to navigate to the past slides to help yourself out if you get lost

Activity

• Add color, size, alpha, and shape aesthetics to your graph.

• Be bold be brave! Experiment!

• What happens when you add more than one aesthetic?

05:00

How would you make this plot?

all we are doing is making each point blue. Should be simple enough. If you are following along please add color = blue where you think it should go!

ggplot() + 
  geom_point(data = gapminder,
   aes(x = gdpPercap,
    y = lifeExp,
    color = "blue"))

Remember when we do things in aes R will look for a column in our dataset. Importantly R will just looks for things it can do given the scope of the function. So if we put blue into the aes portion of it will plot it. But it will get confused because blue is not a column in our data set.

ggplot() +
  geom_point(data = gapminder,
    aes(x = gdpPercap,
        y = lifeExp),
        color = "blue") 

ggplot will take the color argument inside aes and outside aes. If we specify color inside aes than it will color things by a column in the dataset. If you specify color outside of aes ggplot will make everything that color

Same options different stuff

there we can show the same relationship between two variables in a ton of different ways. there are a ton of geoms that we can use to visualize our data. In this case I am just fitting a loess line through it.

What Comes With ggplot

 [1] "geom_abline"            "geom_area"              "geom_bar"              
 [4] "geom_bin_2d"            "geom_bin2d"             "geom_blank"            
 [7] "geom_boxplot"           "geom_col"               "geom_contour"          
[10] "geom_contour_filled"    "geom_count"             "geom_crossbar"         
[13] "geom_curve"             "geom_density"           "geom_density_2d"       
[16] "geom_density_2d_filled" "geom_density2d"         "geom_density2d_filled" 
[19] "geom_dotplot"           "geom_errorbar"          "geom_errorbarh"        
[22] "geom_freqpoly"          "geom_function"          "geom_hex"              
[25] "geom_histogram"         "geom_hline"             "geom_jitter"           
[28] "geom_label"             "geom_line"              "geom_linerange"        
[31] "geom_map"               "geom_path"              "geom_point"            
[34] "geom_pointrange"        "geom_polygon"           "geom_qq"               
[37] "geom_qq_line"           "geom_quantile"          "geom_raster"           
[40] "geom_rect"              "geom_ribbon"            "geom_rug"              
[43] "geom_segment"           "geom_sf"                "geom_sf_label"         
[46] "geom_sf_text"           "geom_smooth"            "geom_spoke"            
[49] "geom_step"              "geom_text"              "geom_tile"             
[52] "geom_violin"            "geom_vline"            

There are many more that we can use too. There are tons of different kinds of plots that we can find for our specific plotting needs that people have written for R I have like an unhealthy obsession with ggridges. I will say geom_sf only gets added when you have the sf package loaded and then

via GIPHY

It’s not that certain geoms should never be used with certain kinds its that sometimes it doesnt make a whole lot of sense or we aren’t getting the most out of our data visualizations. If it takes the arguments it will plot it.

Example(sort of)

Both times we see that we that ggplot will take the arguments and plot them but we arent really getting a lot of useful information. Boxplots would provide a much provide more info

Your Turn

02:00

take the graph on the left and change it to the plot on the right. use the ggplot cheatsheet.

Answer

ggplot() +
  geom_boxplot(data = gapminder,
  aes(x = continent, 
  y = lifeExp))

Your Turn Again

Hint do not supply a Y value

02:00

Lets make a histogram of lifeExp. In this case we only need to supply one variable to aes. Take exp and look at the distribution through a histogram

ggplot() +
  geom_histogram( data = gapminder, 
    aes(x = lifeExp))

Your Turn

Make This Density Plot filled by continent

02:00

Again do not supply a Y column

ggplot() +
  geom_density( data = gapminder,
  aes(x = lifeExp,
     fill = continent),
     alpha = 0.75)

Complex graph!

Map wealth to the x-axis, health to the y-axis, add points, color by continent, size by population, scale the x-axis with a log. Please get rid of the default theme.

Local

ggplot() +
  geom_point(data = gapminder,
   aes(x = gdpPercap,
       y = lifeExp, 
       color = continent)) + 
  geom_smooth(data = gapminder,
   aes(x = gdpPercap, 
       y = lifeExp, 
      color = continent)) 

Global

ggplot(gapminder,
       aes(x = gdpPercap,
          y = lifeExp, 
          color = continent))  + 
  geom_point() +
  geom_smooth() 

So we have mostly been working inside the individual geoms. However most people do not do this because it can start to get inconvenient quickly. If we work inside each individual geom than the other geom will not know what is going on and ggplot will get confused.

Instead if you pass off our argument to the first layer that the aes argument will inheret the appropriate stuff

Building Plots

Starting with Data and aesthics

ggplot(gapminder,
 aes(x = gdpPercap,
    y = lifeExp))

So far we have only walked through geoms and aesthics but we need other layers to make graphs that make sense. And are presentable to professional audiences

Add geom_point

ggplot(gapminder,
    aes(x = gdpPercap,
        y = lifeExp,
        color = continent)) +
  geom_point()

Add geom_smooth

ggplot(gapminder,
    aes(x = gdpPercap,
       y = lifeExp,
       color = continent)) +
  geom_point() +
  geom_smooth()

Change Transparency

ggplot(gapminder,
    aes(x = gdpPercap,
       y = lifeExp,
       color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth()

Adjust scales with scale_x_log10

ggplot(gapminder,
     aes(x = gdpPercap,
         y = lifeExp,
         color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10()

Add axis labels and title with labs

ggplot(gapminder,
      aes(x = gdpPercap,
          y = lifeExp,
          color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10() +
  labs(x = "GDP per cap",
       y = "Life Expectancy",
       title = "The Effect of GDP per cap on Life Expectancy")

Labs can handle a lot of arguments. there are individual functuions for the x label and y label and title. But make your like easier by just using labs. You can also modify the legend in labs but I have various success with that. I tend to modify the legend using guides or theme

Add viridis color scale

ggplot(gapminder,
      aes(x = gdpPercap,
          y = lifeExp,
          color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10() +
  labs(x = "GDP per cap",
       y = "Life Expectanty",
       title = "The Effect of GDP per cap on Life Expectancy") +
  scale_color_viridis_d()

Add theme

ggplot(gapminder,
      aes(x = gdpPercap,
         y = lifeExp,
         color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10() +
  labs(x = "GDP per cap",
       y = "Life Expectanty",
       title = "The Effect of GDP per cap on Life Expectancy") +
  scale_color_viridis_d() +
  theme_bw()

Facet by Continent

ggplot(gapminder,
    aes(x = gdpPercap,
        y = lifeExp,
        color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10() +
  labs(x = "GDP per cap",
      y = "Life Expectanty",
      title = "The Effect of GDP per cap on Life Expectancy") +
  scale_color_viridis_d() +
  theme_bw() +
  facet_wrap(vars(continent))

Change Theme Options

ggplot(gapminder,
  aes(x = gdpPercap,
      y = lifeExp,
      color = continent)) +
  geom_point(alpha = 0.5) +
  geom_smooth() +
  scale_x_log10() +
  labs(x = "GDP per cap",
      y = "Life Expectanty",
      title = "The Effect of GDP per cap on Life Expectancy") +
  scale_color_viridis_d() +
  theme_bw() +
  facet_wrap(vars(continent)) +
  theme(legend.position = "none")

Scales

Example layer	What it does
scale_x_continuous()	Make the x-axis continuous
scale_x_continuous(breaks = 1:5)	Manually specify axis ticks
scale_x_log10()	Log the x-axis
scale_color_gradient()	Use a gradient
scale_fill_viridis_d()	Fill with discrete viridis colors

Often time we need to or want to adjust the scales that one or more of the axis are on or we pass off color arguments to

Scales in Action

ggplot(gapminder,
  aes(x = gdpPercap,
      y = lifeExp,
      size = pop,
      color = continent)) +
  geom_point(alpha = 0.5) +
  labs(x = "Income",
     y = "Life Expectancy") +
  scale_x_log10(labels = scales::dollar) +
  theme_bw() 

Instead of directly transforming your data ggplot allows you to do this through the scale_x function which is pretty neat. I will not go through each of them but that is pretty handy. If you want nicer looking labels you can use the scales package to make nicer looking labels. I.e adding commas to values on the x or y axis adding the dollar sign or percentage sign etc

Scales in Action

ggplot(gapminder,
  aes(x = gdpPercap,
     y = lifeExp,
     size = pop,
    color = continent)) +
  geom_point(alpha = 0.5) +
  labs(x = "Income", y = "Life Expectancy") +
  scale_x_log10(labels = scales::dollar) +
  scale_color_manual(values = c("#04a3bd",
     "#f0be3d",
     "#931e18",
     "#da7901")) +
  theme_bw() 

If you want to use a different set of colors you choose those manull

Scales in Action

ggplot(gapminder,
  aes(x = gdpPercap,
     y = lifeExp,
     size = pop,
    color = continent)) +
  geom_point(alpha = 0.5) +
  labs(x = "Income", y = "Life Expectancy") +
  scale_x_log10(labels = scales::dollar) +
  scale_color_met_d(name = "Veronese") +
  theme_bw() 

Scales

The scale_*() components control the properties of all the
aesthetic dimensions mapped to the data.

The extensions (*) can be filled by e.g.:

• continuous(), discrete(), reverse(), log10(), sqrt(), date() for positions

• continuous(), discrete(), manual(), gradient(), gradient2(), brewer() for colors

• continuous(), discrete(), manual(), ordinal(), area(), date() for sizes

• continuous(), discrete(), manual(), ordinal() for shapes

• continuous(), discrete(), manual(), ordinal(), date() for transparency

Illustration by Allison Horst

Continuous vs. Discrete in {ggplot2}

Continuous:
quantitative or numerical data

• height
• weight
• age
• counts

Discrete:
qualitative or categorical data

• species
• sex
• study sites
• age group

Continuous vs. Discrete in {ggplot2}

Continuous:
quantitative or numerical data

• height (continuous)
• weight (continuous)
• age (continuous or discrete)
• counts (discrete)

Discrete:
qualitative or categorical data

• species (nominal)
• sex (nominal)
• study site (nominal or ordinal)
• age group (ordinal)

Scales in Action

ggplot(gapminder,
      aes(x = gdpPercap,
          y = lifeExp,
          size = pop,
          color = continent)) +
  geom_point(alpha = 0.5) +
  scale_x_continuous(limits = c(0, 30000)) +
  theme_bw()

In some cases you may want to change the coordinates of the plot to expand or contract the coordinates or in some cases just flip them without hassle. ggplot defaults to using coord_cartesian on its own but coord_cartesian can be used to zoom in on part of the plot. The coor_flip here is a bit silly but it is insanely useful for barcharts and other charts where you may have added things into aes and don’t want to go back and redo it

Coordinate Systems

= interpret the position aesthetics

• linear coordinate systems: preserve the geometrical shapes
  • coord_cartesian()
  • coord_fixed()
  • coord_flip()
• non-linear coordinate systems: likely change the geometrical shapes
  • coord_polar()
  • coord_map() and coord_sf()
  • coord_trans()

Change the Limits of the plot

ggplot(gapminder,
    aes(x = gdpPercap,
     y = lifeExp)) +
  geom_point(alpha = 0.5) +
  scale_x_continuous(limits = c(0, 30000)) +
  theme_bw()

Circular Coordinate Systems

ggplot(penguins,
 aes(x = species,
 fill = species)) +
geom_bar() + 
coord_polar()

ggplot(penguins,
 aes(x = species,
  fill = species)) +
geom_bar() + 
coord_cartesian()

Your Turn

Change the colors of this density plot

04:00

look at the help file for scale_fill_viridis and play around with the options

How I Did It

ggplot(penguins,
 aes(x = bill_length_mm,
    fill = species)) +
  geom_density( alpha = 0.75) +
  theme_bw() +
  scale_fill_viridis_d(option = "magma")

Facets

Example layer	What it does
facet_wrap(vars(continent))	Plot for each continent
facet_wrap(vars(continent, year))	Plot for each continent/year
facet_wrap(…, ncol = 1)	Put all facets in one column
facet_wrap(…, nrow = 1)	Put all facets in one row

facet_wrap

ggplot(gapminder,
  aes(x = gdpPercap,
     y = lifeExp,
     size = pop)) +
  geom_point(alpha = 0.5) +
  theme_bw() +
  scale_x_log10() +
  facet_wrap(vars(continent)) 

If we want to get individual plots for each value of the variable we can feed this to facet wrap and is a really useful way of displaying information and saves you from creating subsets for each continent than plotting them individuall than combining them. You also have lots of

facet_grid

ggplot(data = filter(gapminder,
 year %in% c(1987,1997,2002, 2007)),
    aes(x = gdpPercap,
     y = lifeExp,
     size = pop)) +
  geom_point(alpha = 0.5) +
  theme_bw() +
  scale_x_log10() +
  facet_grid(vars(year))

forms a matrix of panels defined by row and column faceting variables. It is most useful when you have two discrete variables, and all combinations of the variables exist in the data.

facet_grid

ggplot(data = filter(gapminder,
 year %in% c(1987,1997,2002, 2007)),
    aes(x = gdpPercap,
     y = lifeExp,
     size = pop)) +
  geom_point(alpha = 0.5) +
  theme_bw() +
  scale_x_log10() +
  facet_grid(vars(year), vars(continent))

Labels

Example layer	What it does
labs(title = “Neat title”)	Title
labs(caption = “Something”)	Caption
labs(y = “Something”)	y-axis
labs(size = “Population”)	Title of size legend

Labels with labs

ggplot(gapminder, 
       aes(x = gdpPercap,
        y = lifeExp, 
        color = continent,
        size = pop)) +
 geom_point(alpha = 0.5) +
  scale_x_log10() +
  labs(title = "Health and wealth grow together",
       subtitle = "Data from 2007",
       x = "Wealth (GDP per capita)",
       y = "Health (life expectancy)",
       color = "Continent",
       size = "Population",
       caption = "Source: The Gapminder Project")

The labs argument has a ton of flexibility. Basically if you feed something into aes you can change it in labs. I Tend to change legend labels in guides. That is just a personal preference but labs is superflexible. You will sometimes see other solutions for changing your legend.

Changing the Default Theme

theme_minimal

theme_dark

The theme argument

• Has lots and lots of options(94 to be exact)

• You can change basically anything you could think of in a plot

  • My ggplot theme is basically just a some tweaks to theme arguments

theme_bw() + 
theme(legend.position = "bottom",
      plot.title = element_text(face = "bold"),
      axis.title.y = element_text(face = "italic"))

Saving your work

your_plot_here = ggplot(data, aes(x = blah, y = blah))

ggsave("name-of-your-file.pdf",your_plot_here) 

ggsave("name-of-your-file.pngs",your_plot_here)

Making Maps

New Packages

install.packages(c("sf", "tidygeocoder"))
devtools::install_github("ropenscilabs/rnaturalearth")
library(rnaturalearth)
library(sf)

• If you are on a Mac please go to the r-spatial-website if you run into problems

• The workhorse for this particular section will be sf

There are lots of mapping packages in R. I will only show you how to map in ggplot. You can get a really passable map in ggplot. I have included the code the Emre wrote that uses tmap. tmap is good and I have used it but once you get the hand of ggplot I think it is just easier to use ggplot than to learn how to use a new package

There are some other packages like geom_map and ggmap. They still work but sf(simple features) is really great and has a ton of support and resources. See this book

Mapping in R

• R and ggplot can get you pretty far

• The stuff from these workshops broadly apply

  • including your dplyr verbs

• Lots of your needs to make static maps can be met

• Depending on what you are doing you may have to wait a bit

Map Made By Kieran Healy

Shape Files

• To read in shape files you use read_sf you should see something that looks like this!

NAME	geometry
Mordor	POINT (1330373 596482.5)
Hobbiton	POINT (515948 1043820)
Edoras	POINT (853993.3 723854.1)
Rivendell	POINT (884331.5 1057787)
Minas Tirith	POINT (1111425 621234.6)

When you download a shapefile you will get lots and lots of stuff. Lots of the shapefiles that you encounter in the wild are maintained by governments so keep those in the working directory. I will get to how to read in shape files in a minute. But when you read them into R the only thing that changes from your usual dataframe, tibble, data.table etc is that you will get a column in your dataset called geometry. This is just a list of things that delineate boundaries. for smaller countries like Fiji the number of boundaries are going to be smaller than say larger countries by area. It doesnt make sense to us but it makes sense to R

Making Maps in ggplot

## devtools::install_github("ropenscilabs/rnaturalearth")
library(rnaturalearth)

world_map_ne = ne_states(returnclass = "sf")

ggplot() +
  geom_sf(data = world_map_ne)

as you can see nothing really fundamentally changes when you are making maps in ggplot. It is no different fundamentally than making a scatter plot or any other kind of plot here we are just feeding data to geom_sf

Changing the Projections

ggplot() +
  geom_sf(data = world_map_ne) + 
  coord_sf(crs = "+proj=cea +lon_0=0 +lat_ts=45") +
  theme_void()

You can really easily change the projections you want. Here we are just feeding it a different projections If you ever watched the west wing you may recognize this particular projection

Working Without Shape Files(kind of)

• As is the case sometimes we do not have a shape file

• Don’t worry sf has you covered

• You just need to feed it the right things

Making a Bespoke Shapefile

• You can either feed it latitude and longitudes

• Or you can feed it addresses

• Most free ones have rate limits

  • So be mindful of the size of your data

In this case if we do not have the geometry column but we have some addresses we can just geocode them to produce the column.

Making a Bespoke Shapefile(cont)

ga_cities = tribble( 
  ~city, ~lat, ~long,
  "Atlanta", 33.748955, -84.388099,
  "Athens", 33.950794, -83.358884,
  "Savannah", 32.113192, -81.089350
)


ga_cities_geometry = ga_cities |>  
  st_as_sf(coords = c("long", "lat"), crs = st_crs("EPSG:4326"))
ga_cities_geometry

city	geometry
Atlanta	POINT (-84.3881 33.74896)
Athens	POINT (-83.35888 33.95079)
Savannah	POINT (-81.08935 32.11319)

examples derived from Andrew Heiss

Making a Bespoke Shapefile(cont)

library(tidygeocoder)
breweries_I_visit = tribble(
  ~name, ~address,
  "Russian River Brewing", "725 4th St, Santa Rosa, CA 95404",
  "Orpheus Brewing", "1440 Dutch Valley Pl NE, Atlanta, GA 30324",
  "Three Tavens", "121 New St, Decatur, GA 30030",
  "HenHouse Brewing Company", "322 Bellevue Ave, Santa Rosa, CA 95407"
)


breweries_geocode = breweries_I_visit |> 
  geocode(address, method = "osm")

breweries_geocode |> 
  st_as_sf(coords = c("long", "lat"), crs = st_crs("EPSG:4326")) |> 
  knitr::kable(format = "html")

name	address	geometry
Russian River Brewing	725 4th St, Santa Rosa, CA 95404	POINT (-122.7117 38.4418)
Orpheus Brewing	1440 Dutch Valley Pl NE, Atlanta, GA 30324	POINT (-84.36874 33.79355)
Three Tavens	121 New St, Decatur, GA 30030	POINT (-84.28488 33.77306)
HenHouse Brewing Company	322 Bellevue Ave, Santa Rosa, CA 95407	POINT (-122.7255 38.40122)

in the geocode argument you can cascade the services just in case it fails. The neat thing is that you can feed it a dataframe of addresses. What is also pretty cool is that it plays really well with other R tidyverse packages so if you scrape a website to grab a ton of addresses than you can just feed it into tidygeocoder

from there you can use ggplot or even create interactive leaflet maps

Mapping Middle Earth

# using some programming read in everything iteratively
# there is probably a simpler solution but this worked for me 
shapes <-  list.files(path = "data/ME-GIS/", pattern = "*.shp", full.names = TRUE)

shapes_read = map(shapes, read_sf)

remove <- c("data/ME-GIS//", ".shp", "2", "02", "_18", "")
#https://stackoverflow.com/questions/29036960/remove-multiple-patterns-from-text-vector-r
shapes_names = shapes |> 
  str_remove_all(paste(remove, collapse = "|")) |> 
  str_to_lower()

names(shapes_read) = shapes_names

map(names(shapes_read), ~assign(.x, shapes_read[[.x]], envir = .GlobalEnv))

places = placenames |>
  filter(NAME %in% c("Hobbiton",
                     "Rivendell",
                     "Edoras",
                     "Minas Tirith"))

mordor = placenames[placenames$NAME == "Mordor",]

mountains_to_label = mountains_anno[mountains_anno$name == "Erebor The Lonely Mountain",]


ggplot() +
  geom_sf(data = contours,
          size = 0.15,
          color = "grey90") +
  geom_sf(data = coastline,
          size = 0.25,
          color = "grey50") +
  geom_sf(data = rivers,
          size = 0.2,
          color = "#0776e0",
          alpha = 0.5) +
  geom_sf(data = lakes,
          size = 0.2,
          color = "#0776e0",
          fill = "#0776e0") +
  geom_sf(data = forests,
          size = 0,
          fill = "#035711",
          alpha = 0.5) +
  geom_sf(data = mountains_anno, size = 0.25) +
  geom_sf(data = places) +
  geom_sf_label(data = filter(places, NAME !="Rivendell"),
                aes(label = NAME),
                nudge_y = 80000, size = 6) +
  geom_sf_label(data = filter(places, NAME == "Rivendell"),
                aes(label = NAME),
                nudge_x = 88000, size = 6) +
  geom_sf_label(data = mountains_to_label,
                aes(label = name),
                nudge_y = 80000, size = 6) +
  geom_sf_label(data = mordor,
                aes(label = NAME),
                nudge_y = -10000, size = 6) +
  theme_void() +
  theme(plot.background = element_rect(fill = "#fffce3"))

As you can see this does not look fundamentally all that different from the rest of our ggplot code. We are just coloring and sizing ssutff depening on what we want. We can add annotations