NCCU R Intro Carpentries Workshop Notes

script.R

########################################
### Intro to R NCCU, Aug 12-14, 2020 ###
########################################

# this is the workshop script

# get working directory:
getwd() # ctrl + enter to execute this line

# just a test shift + ctrl + c > make a comment (more shortcuts under Tools > shortcuts)
# another line 
# and anogtehr

# download data
# create a new data directory
dir.create("data")

# create two more directories: data_output, fig_output
dir.create("data_output")
dir.create("fig_output")

# download the dataset from figshare into your local data folder
download.file("https://ndownloader.figshare.com/files/11492171", "data/SAFI_clean.csv", mode = "wb")

# install additional packages
install.packages("tidyverse", dependencies = TRUE)

######

# creating objects
3 + 8
12 / 7

area_hectares <- 1.0  # assigns a value to object
area_hectares = 1.0

# alt / opt - <- 

area_hectares # retrieve value of object

# use object in calculation
2.478 * area_hectares

area_hectares <- 2.5
area_hectares

area_acres <- 2.47 * area_hectares
area_hectares <- 50


# from eterhpad
length <- 19
width <- 12
area = length * width
area
length <- 50 # change value 
area # value not changed
area = length * width # multply again!
area # value changed


# functions in R
getwd() # no input argument
a <- 64
sqrt(a)
b <- sqrt(a)
b

round(3.141593)
# return 2 digits
round(3.141593, digits = 2)
?round

round(3.141593, 2) # keep order of arguments as defined 
round(digits = 2, x = 3.141593) # name arguments to switch order!

################################################

# Vectors

hh_members <- c(3, 7, 10, 6)
hh_members # what is in the vector?

# a vector of characters
respondent_wall_type <- c("muddaub", "burntbricks", "sunbricks")
respondent_wall_type

length(hh_members)
length(respondent_wall_type)

# what type is the element?
class(hh_members)
class(respondent_wall_type)

# structure of object?
str(hh_members)
str(respondent_wall_type)

# add values to a vector
possessions <- c("bicycle", "radio", "television")
possessions <- c(possessions, "mobile_phone")
possessions # it now has 4 elements!

# object names in R
2x <- 4 # cannot begin with numbers
for # don't use reseved words
length <- 4 # avoid using function names

possessions <- c("car", possessions)
possessions 

# other data types
# logical
my_logical <- c(TRUE, FALSE, FALSE)
class(my_logical)

# integer
my_integer <- c(2L, 4L, 4L)
class(my_integer)


# from exercise 2
# what class are those?
num_char <- c(1, 2, 3, "a")
class(num_char) # a character vector
num_char
num_logical <- c(1, 2, 3, TRUE) 
class(num_logical) # a numeric vector
num_logical
char_logical <- c("a", "b", "c", TRUE)
class(char_logical)
char_logical
tricky <- c(1, 2, 3, "4")
class(tricky)
tricky
# 1. a vector can only be of one element type
# 2. character will override  numeric will override logical

num_logical <- c(1, 2, 3, TRUE)
num_logical
char_logical <- c("a", "b", "c", TRUE)
char_logical
combined_logical <- c(num_logical, char_logical)
combined_logical

# subsetting
respondent_wall_type

# by index:
respondent_wall_type[2] # extract 2nd element
respondent_wall_type[-2] # everything except 2nd element

respondent_wall_type[c(2, 3)] # extract 2nd and 3rd element
respondent_wall_type[-c(2, 3)] # extract all except 2nd and 3rd element

respondent_wall_type[c(1, 2, 3, 1, 3 ,3, 3)]
respondent_wall_type

respondent_wall_type <- respondent_wall_type[c(1, 3, 3, 2)]
respondent_wall_type


# conditional subsetting
hh_members
hh_members[c(TRUE, FALSE, TRUE, TRUE)]

# example condition
hh_members > 5
hh_members[hh_members > 5]

hh_members[hh_members < 3 | hh_members > 5] # | means OR
hh_members >= 7 # greater or equal
hh_members == 7 # equal
hh_members != 7 # not equal
hh_members > 3 & hh_members < 10   # & means AND

possessions
possessions[possessions == "car"]
possessions[possessions == "car" | possessions == "bicycle"]

possessions %in% c("car", "bicycle")
possessions[possessions %in% c("car", "bicycle")]


# from exercise 3a/b
# extract two lowest values 
hh_income <- c(177, 134, 172, 157, 100)


hh_income[hh_income<135]

length(hh_income)
hh_income[hh_income <= 100 | hh_income <= 150 ]
hh_income[hh_income <= 150]

hh_income [hh_income<=135]


# extract "television"
possessions <- c("car", "bicycle", "radio", "television", "mobile_phone") 

possessions [possessions=="television"]
possessions[4]
possessions[possessions == "television"]

# missing data
rooms <- c(2, 1, 1, NA, 4)
max(rooms)
max(rooms, na.rm = TRUE)
mean(rooms)
mean(rooms, na.rm = TRUE)

# extract which elements are missing
is.na(rooms)
# extract elemets that are NOT missing
!is.na(rooms)
rooms[!is.na(rooms)]

# how many elements are NA
is.na(rooms)
sum(is.na(rooms)) # one element is NA!

complete.cases(rooms) # returns a locigal vector
rooms[complete.cases(rooms)]

na.omit(rooms) # returns object with NA removed


# from exercise 4
rooms <- c(1, 2, 1, 1, NA, 3, 1, 3, 2, 1, 1, 8, 3, 1, NA, 1)
# median
median(rooms[complete.cases(rooms)])
median(rooms,na.rm = TRUE)

median(rooms[!is.na(rooms)])

median(rooms,na.rm = TRUE)
median(rooms[!is.na(rooms)])
rooms <- rooms[na.omit(rooms)]
median(rooms)


# remove NAs
roomscomp <- (rooms[complete.cases(rooms)])
cleaned_rooms <- rooms[!is.na(rooms)]
rooms <- c(rooms[!is.na(rooms)])
rooms <- rooms[na.omit(rooms)]
clean_rooms <- rooms[!is.na(rooms)]
clean_rooms

# households with more than 2 rooms
rooms[rooms >2]
length(rooms[rooms >2])

cleaned_rooms <- rooms[!is.na(rooms)]
cleaned_rooms
length(cleaned_rooms[cleaned_rooms > 2])


####### GGPLOT #########

library(tidyverse)

#  Download link for interviews_plotting.csv
#https://drive.google.com/file/d/1AVDcIcgWeQs4pEQVCaxMzQtdziHCMRYt/view?usp=sharing

# Unzip and place in your data_output folder.

interviews_plotting <- read_csv("data_output/interviews_plotting.csv")


# scatter plots
ggplot(interviews_plotting, # data
       aes(x = no_membrs, y = number_items)) + # mapping - what to plot?
  geom_point() # geometry, visual properties, how to display it?

ggplot(interviews_plotting, aes( x = no_membrs, y = number_items)) +
  geom_point(alpha = 0.5)

ggplot(interviews_plotting, aes( x = no_membrs, y = number_items)) +
  geom_jitter(alpha = 0.5)

ggplot(interviews_plotting, aes( x = no_membrs, y = number_items)) +
  geom_jitter(alpha = 0.5, color = "blue")

ggplot(interviews_plotting, aes( x = no_membrs, y = number_items)) +
  geom_jitter(aes(color = village), alpha = 0.5)


interviews_plot <- ggplot(interviews_plotting, aes( x = no_membrs, y = number_items))

interviews_plot + 
  geom_jitter()

# barplots

ggplot(interviews_plotting, aes(x = respondent_wall_type)) +
  geom_bar()   # number of cases

ggplot(interviews_plotting, aes(x = respondent_wall_type)) +
  geom_bar(aes(fill = village))   # number of cases colored by village

ggplot(interviews_plotting, aes(x = respondent_wall_type)) +
  geom_bar(aes(fill = village), position = "dodge")   # number of cases colored by village

interviews_plotting %>% 
  filter(respondent_wall_type != "cement") %>% 
  count(village, respondent_wall_type) %>% 
  group_by(village) %>% 
  mutate(percent = n / sum(n) * 100) %>% 
  ggplot(aes(x = village, y  = percent, fill = respondent_wall_type)) +
  geom_col(position = "dodge") + # values in the data (instead of counts) 
  labs(title = " my great plot", x = "Village", y = "Percent") +
  theme_bw() +
  theme(text = element_text(size = 16))