Skip to content

Latest commit

 

History

History
336 lines (215 loc) · 7.55 KB

session_1.md

File metadata and controls

336 lines (215 loc) · 7.55 KB
Raw

C4TS R-Club - Session 1

Jason Pott 17/06/2019

## Registered S3 method overwritten by 'rvest':
##   method            from
##   read_xml.response xml2

C4TS R-Club

In this session: - We made sure that our Rstudio setup was working

  • we created a new R-markdown document a handy cheat sheet can be found here

  • We customised our layout of Rstudio panes (Menu > View > Panes > Pane Layout)

Suggested Pane Layout

  • We started to use ggplot for graphing

  • We learnt how to bring data into R

R markdown combined text that is formatted in “markdown”

Rmarkdown can be output to many formats with the defaults being html, pdf and word. Addons are available for power point and other formats.

Installing packages

The R language includes a huge number of functions and syntax out of the box. These “base R” functions can be added to through the installation of other packages.

To install al package type install.packages(“nameofthepackage”) packages must be installed once for each machine they can then be called into active memory by loading them using the command:

library(nameofthepackage)

Note the difference between the functions in the install.packages() function the nameofthepackage is between quotation marks both ‘nameofthepackage’ or “nameofthepackage” are valid.

Quotation marks are not needed once the package has been installed.

Install the tidyverse package you can read more about what the tidyverse is here

to run commands in Rmarkdown move the cursor to the line you want torun and press Command + Enter (on Mac) or Ctrl + Enter (on windows)

ggplot

The next section touches on some of the concepts presented in R for data science - Chapter 3

We are going to creat a graph using the ggplot package. We create a ggplot object by calling ggplot and tell the function what data to use.

mtcars is a data set included in the tidyverse type ?mtcars in R to find out more about it.

ggplot(data = mtcars)

The code creates a blank box. This is the basis of a ggplot.

We can add axis to the plot using the following terms

ggplot(data = mtcars, mapping = aes(x = mpg, y = hp))

Then we can add a visualisation of the data in question by using a geom. Note that to do this, we chain a new line onto the ggplot function using the + symbol

ggplot(data=mtcars, mapping = aes(x = mpg, y = hp)) + 
  geom_point()

We can add labels and annotations using the labs function

ggplot(data=mtcars, mapping = aes(x = disp, y = hp)) + 
  geom_point() +
  labs(title = "title", y = "y", x = "x", subtitle = "subtitle", tag = "tag")

And some color using the color/colour variable of mapping

ggplot(data=mtcars, mapping = aes(x = disp, y = hp, color = wt)) + 
  geom_point() +
  labs(title = "title", y = "y", x = "x", subtitle = "subtitle", tag = "tag")

There are numerous types of Geom which can be used in ggplot a full list can be seen on the ggplot reference pages

Each geom has default statistical transformation included as a default behaviour.

Using the diamonds data set we can produce a histogram of the number of diamonds by cut.

ggplot(data = diamonds, mapping = aes(x = cut)) +
  geom_bar()

Note that in the above graph we have not given R any information about the number of each diamond. Just that we want a bar chart where the x axis is cut and the data set was diamonds.

The default statistic used by geom_bar is count, read more about geom_bar and any R function using ?functionname i.e. {r ?geom_bar

If we wanted to make the same graph using data where we had already calculated the counts of variables we would need to change the way we use geom_bar.

In the next code chunk I have calculated a tally of diamonds by their cut.

test <- diamonds %>% 
  group_by(cut) %>% 
  tally()


# The next function just presents the data in a table
kable(test, "html") %>%
  kable_styling(bootstrap_options = "striped", full_width = F)

cut

n

Fair

1610

Good

4906

Very Good

12082

Premium

13791

Ideal

21551

if we pass this to ggplot as we did before we get a graph that looks nothing like what we were expecting.

ggplot(data = test, mapping = aes(x = cut)) +
  geom_bar()

This is because the stat for geom_bar has used count and so counted the number of rows where each cut type can be found. As each cut type appears only once the resulting value is one for each cut.

The correct way to pass the values to ggplot for the formating of the graph is shown below but importantly stat = “identity” is used in the function call to geom_bar to tell it to use the values in n and not calculate these.

# I am trying to show the default statistics functions of geoms
ggplot(data = test, mapping = aes(x = cut, y=n)) +
  geom_bar(stat = "identity")

Loading data into R

Loading data into R requires a consistent approach.

you need to know the precise location of a file path (the location where a file is stored) and you need to give this to R in a way that it can understand based on the current working directory.

getwd()
## [1] "/Users/jason/R/r4ds/01. Session 1 "

getwd() returns the current working directory where R is being used. When working in Rmarkdown documents it will typically be where ever the rmarkdown document is saved.

To make things easy when starting out create a new project folder for each project that you are working on and keep your data files in the same location.

The read.csv function can be used to read in files in csv format and it comes from the readr package which is part of the tidyverse.

read.csv(file = "test.csv", header = FALSE) -> data
## Warning in read.table(file = file, header = header, sep = sep, quote =
## quote, : incomplete final line found by readTableHeader on 'test.csv'

print(data)
##   V1 V2
## 1  1  a
## 2  2  b
## 3  3  c

Next session we will be covering: - [ ] working in projects
- [ ] loading data files
- [ ] data calculations
- [ ] graph something
- [ ] create a table