dplyr and ggplot2.First steps.
readr, dplyr, and ggplot2library(dplyr)
Attaching package: 'dplyr'The following objects are masked from 'package:stats':
filter, lagThe following objects are masked from 'package:base':
intersect, setdiff, setequal, unionlibrary(readr)
library(ggplot2)
library(glue) # just to make long strings wrap in PDFThese are the three libraries you need most when you explore a tabular dataset.
myfilepath <-"datasets_ATRIUM/gapminder_hourly_labour_cost_constant_2017_usd--by--geo--time.csv"
laborcost_df <- read_csv(file = myfilepath,
show_col_types = TRUE)Rows: 548 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): geo
dbl (2): time, hourly_labour_cost_constant_2017_usd
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.It is the file demonstrated in the previous session. We saved it into the folder datasets_ATRIUM. You can also get it at https://raw.githubusercontent.com/open-numbers/ddf--gapminder--systema_globalis/refs/heads/master/countries-etc-datapoints/ddf--datapoints--hourly_labour_cost_constant_2017_usd--by--geo--time.csv.
Watch the message read_csv gives you about the file. You can suppress it by overriding the default to show_col_types = FALSE.
dplyr::glimpseglimpse(laborcost_df)Rows: 548
Columns: 3
$ geo <chr> "arg", "arg", "arm", "arm", "arm"…
$ time <dbl> 2011, 2012, 2011, 2012, 2013, 201…
$ hourly_labour_cost_constant_2017_usd <dbl> 0.92, 1.04, 4.23, 4.59, 6.12, 6.0…Gives you the number of rows and columns, the column names with their data class, and it also shows as many elements (values) in each column as to fit your screen.
summarysummary(laborcost_df) geo time hourly_labour_cost_constant_2017_usd
Length:548 Min. :1994 Min. : 0.000
Class :character 1st Qu.:2005 1st Qu.: 9.867
Mode :character Median :2011 Median :18.320
Mean :2010 Mean :19.686
3rd Qu.:2017 3rd Qu.:26.915
Max. :2020 Max. :48.720 Gives you the “five-number summary” of each numeric column (it’s often called this way, although the numbers are obviously six…). With categorical columns, it depends, whether the column is a character vector or a factor.
summary with categorical columns as factors geo time hourly_labour_cost_constant_2017_usd
cze : 22 Min. :1994 Min. : 0.000
svn : 22 1st Qu.:2005 1st Qu.: 9.867
cyp : 21 Median :2011 Median :18.320
deu : 21 Mean :2010 Mean :19.686
pol : 21 3rd Qu.:2017 3rd Qu.:26.915
svk : 21 Max. :2020 Max. :48.720
(Other):420 If you have a data frame with categorical variables converted to factors, the summary will show you a glimpse of their levels (unique values) and their frequencies, as well as tell you how many levels there are.
So far, do not worry about factors. The dplyr as well as the ggplot2 libraries do this factor conversion on the fly whenever they need it.
hourly_labour_cost_constant_2017_usd too long, shorten to labor_cost.
colnames(laborcost_df)[colnames(laborcost_df) ==
"hourly_labour_cost_constant_2017_usd"] <- "labor_cost"colnames(laborcost_df)[1] "geo" "time" "labor_cost"dplyrlaborcost_df <- rename(.data = laborcost_df,
labor_cost = hourly_labour_cost_constant_2017_usd
)
colnames(laborcost_df)[1] "geo" "time" "labor_cost"You already know you could have named all columns your way when reading in the file. Here are two ways to rename a column: one base-R-like and the other one provided by dplyr.
dplyrcze_deu_df <- dplyr::filter(laborcost_df, geo %in% c("cze", "deu"))
write_csv(cze_deu_df, "datasets_ATRIUM/gapminder_laborcost_cze_deu.csv")dplyr::filter keeps only rows that meet a condition (or a set of conditions) that you determine with logical operators.
We will return to this in more detail in a separate session.
dplyr::distinctdplyr::distinct(.data = laborcost_df, geo, .keep_all = FALSE)| geo |
|---|
| arg |
| arm |
| aus |
| aut |
| aze |
| bel |
| bgr |
| can |
| che |
| chl |
| cri |
| cyp |
| cze |
| deu |
| dnk |
| esp |
| est |
| fin |
| fra |
| gbr |
| geo |
| grc |
| hrv |
| hun |
| irl |
| isl |
| isr |
| ita |
| kaz |
| ltu |
| lux |
| lva |
| mda |
| mkd |
| mlt |
| mus |
| nld |
| nor |
| nzl |
| phl |
| pol |
| prt |
| rou |
| rus |
| svk |
| svn |
| swe |
| ukr |
The data set contains almost 50 countries. They will probably not be distinguishable in the plot.
laborcost_plot <- ggplot(data = laborcost_df,
mapping = aes(x = time,
y = labor_cost,
color = geo)) +
geom_point()
laborcost_plot
laborcost_plot_size <- ggplot(data = laborcost_df,
mapping = aes(x = geo,
y = time,
alpha = labor_cost)) +
geom_point()
laborcost_plot_size
How does it capture the variables now? Is it telling a different story?
ggsave(filename = "my_output_files/laborcost_plot.svg",
plot = laborcost_plot)Saving 7 x 5 in imageggsave(filename = "my_output_files/laborcost_plot.png",
plot = laborcost_plot,
device = grDevices::png) Saving 7 x 5 in image# device = "png" or this when RStudio hickups
ggsave(filename = "my_output_files/laborcost_plot.pdf",
plot = laborcost_plot)Saving 7 x 5 in imagelist.files(path = "my_output_files/", pattern = "laborcost_plot")[1] "laborcost_plot.pdf" "laborcost_plot.png" "laborcost_plot.svg"ggsave wants a file name (with path) including the format suffix.
Sometimes, especially when you run your scripts over and over in Quarto, something breaks behind the scenes and you get cryptical error messages about mismatch of graphical devices or similar. Sometimes the best solution is to restart R, close the file, clean the Environment and reopen the file. Or run the command directly in the Console.
ggplot2Specific syntax
A plot is an object (goes in a variable)
Can be saved to files - include desired format in the file name
Maps variables on X, Y, color, transparency…
Feel free to add more insights or impressions.
ggplot2 \(\approx\) implemented Grammar of Graphicshttps://ggplot2.tidyverse.org/
All plots have the same logic and components in a few layers.
Not just drawings, statistical transformations behind the scenes (e.g. histogram)
When you see a ggplot2 plot you have an idea how the source table is structured
Grammar of Graphics is a book by Leland Wilkinson.
Base R plots: many independent packages \(\rightarrow\) less consistence, harder to learn properly
You cannot save base R plots as objects.
ggplot2Data
Aesthetic mappings + Facets (subgraphs)
Geometric objects (aka geoms)
Statistical transformations (aka stats)
Coordinate system
Theme
Today just data, aesthetic mappings and geoms. This is enough to make an informative plot if you don’t need it super pretty for publication.
data frame with tidy data structure
each observation on one row
each variable in one column
categorical variables automatically read as factors
axes X, Y
shape / linetype
color / fill / stroke
size / linewidth
alpha (transparency)
label
plot types, such as:
histogram
scatterplot
barplot
boxplot
heatmap
and many others
Try and sketch these plots by hand (fake data). What are these plots good at telling? How many of which variables?
cze_deu_df <- read_csv("datasets_ATRIUM/gapminder_laborcost_cze_deu.csv")Rows: 43 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): geo
dbl (2): time, labor_cost
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.ggplot(data = cze_deu_df,
mapping = aes(x = time, y = labor_cost, color = geo)) +
geom_point(size = 7)
ggplot(data, mappings) + geom_…( )
or
ggplot(data) + geom_…(mappings)
The mapping argument always takes the aes() function.
The plus sign never works at the beginning of a new line.
The geometric object (actual plot) is generated by a geom_something() function.
Different geom_ functions require/accept different aesthetic scales. Look them up in help/cheat sheet.
10.1 Comment on the plot
How do you call plots with points and two axes?
How many variables does the plot capture and how? Which are the types of variables?
Look at the script. Try to dissect it in parts and interpret them.