author: Hazel Kavili date: 2017-09-30 width: 1400 height: 1280 font-family: 'Helvetica'
Let's meet with the most used IDE (integrated desktop environment) for R
- Scripts, Console,
- Code and workflow are more reproducible if we can document everything that we do.
- Environment, History, Connections
- Files, Plots, Packages, Help, Viewer
- The viewer window will helpy ou to see Plots, Shiny applications, blog pages of you did!
- Ask Google
- Search in Stackoverflow
- An introduction to R
- R for Data Science book from Hadley Wickham & Garrett Grolemund
- Try R
- R mailing list: first, learn how to ask questions!
- R Tutorials
- Get help from R:
help.start()
help(mean)
?mean
example(mean)head(mtcars) mpg cyl disp hp drat wt qsec vs am gear carb
Mazda RX4 21.0 6 160 110 3.90 2.620 16.46 0 1 4 4
Mazda RX4 Wag 21.0 6 160 110 3.90 2.875 17.02 0 1 4 4
Datsun 710 22.8 4 108 93 3.85 2.320 18.61 1 1 4 1
Hornet 4 Drive 21.4 6 258 110 3.08 3.215 19.44 1 0 3 1
Hornet Sportabout 18.7 8 360 175 3.15 3.440 17.02 0 0 3 2
Valiant 18.1 6 225 105 2.76 3.460 20.22 1 0 3 1
tail(mtcars) mpg cyl disp hp drat wt qsec vs am gear carb
Porsche 914-2 26.0 4 120.3 91 4.43 2.140 16.7 0 1 5 2
Lotus Europa 30.4 4 95.1 113 3.77 1.513 16.9 1 1 5 2
Ford Pantera L 15.8 8 351.0 264 4.22 3.170 14.5 0 1 5 4
Ferrari Dino 19.7 6 145.0 175 3.62 2.770 15.5 0 1 5 6
Maserati Bora 15.0 8 301.0 335 3.54 3.570 14.6 0 1 5 8
Volvo 142E 21.4 4 121.0 109 4.11 2.780 18.6 1 1 4 2
Base R codes:
mtcars$transmission <-
ifelse(mtcars$am == 0, "automatic", "manual")dplyr codes:
mtcars <- mtcars %>%
mutate(transmission = case_when(am == 0 ~ "automatic", am == 1 ~ "manual"))Base R codes:
mtcars$trans_color <-
ifelse(mtcars$transmission == "automatic", "green", "blue")
pdf("plots/scatter_base.pdf", width = 5, height = 3)
plot(mtcars$mpg ~ mtcars$disp, col = mtcars$trans_color)
legend("topright",
legend = c("automatic", "manual"),
pch = 1, col = c("green", "blue"))
dev.off()ggplot codes:
p1 <- ggplot(mtcars, aes(x = disp, y = mpg, color = transmission)) +
geom_point()
ggsave("plots/scatter_tidy.pdf", p1, width = 5, height = 3)R commands:
- are case sensitive
- can be seperated either by a semi-colon(;), or by a newline
- #comment
Objects:
- varibables, arrays of numbers, character strings, functions
Assignment and Basic Operators
- use <- for assigments
- +,-, *, /, ^, %%
Logical Operators
- <,>, <=, >=, ==, !=, !x, x & y, x | y
Others
- sum, sqrt, min, max, mean, var, sd, abs, summary
- Vectors
- List
- Matrices
- Arrays
- Factors
- Data Frames
- use c() for concatenate more than one element.
- in programming vectors are variable sized sequence of values (not necessarily numbers).
books <- c("history", "sci-fi", "fantasy")
print(books)[1] "history" "sci-fi" "fantasy"
print(class(books))[1] "character"
ages <- c(12,13,14,15,9,8)
print(ages)[1] 12 13 14 15 9 8
print(class(ages))[1] "numeric"
#this is R-Ladies Istanbul
X <- 10
x <- 5
print(paste("X is", X))[1] "X is 10"
print(paste("x is", x))[1] "x is 5"
cat("X and x are equal? = ", X == x)X and x are equal? = FALSE
myNumbers <- c(1:10) # c is short for concatenate
rep(myNumbers, times = 3) [1] 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3
[24] 4 5 6 7 8 9 10
twice <- rep(myNumbers, each = 2)
print(twice) [1] 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10
y <- c(1,2,3,10,15,20)
z <- c(y,4,5,6,y)
print(y)[1] 1 2 3 10 15 20
- Vector Arithmetic
5/y[1] 5.0000000 2.5000000 1.6666667 0.5000000 0.3333333 0.2500000
5*y[1] 5 10 15 50 75 100
y^2[1] 1 4 9 100 225 400
sqrt(y)[1] 1.000000 1.414214 1.732051 3.162278 3.872983 4.472136
seq(from = -10, to = 10, by = 1) [1] -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
[18] 7 8 9 10
seq(from = 1, length = 25, by = 2 ) [1] 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45
[24] 47 49
weights <- c(55, 60, 45, 70, 56, 73, 59, 82)
sum(weights)[1] 500
mean(weights)[1] 62.5
sd(weights)[1] 11.77164
var(weights)[1] 138.5714
length(weights)[1] 8
- Lists can contain many different types of elements inside.
myList <- list(c(1,2,3), 15, "hello")
print(myList)[[1]]
[1] 1 2 3
[[2]]
[1] 15
[[3]]
[1] "hello"
- Select an element from lists
myList[1][[1]]
[1] 1 2 3
myList[[1]][2][1] 2
myList[2][[1]]
[1] 15
- A matrix is two-dimensional recteangular data set.
myMatrix <- matrix(c(1,2,3,4,5,6), nrow = 2, ncol = 3, byrow = TRUE)
print(myMatrix) [,1] [,2] [,3]
[1,] 1 2 3
[2,] 4 5 6
- Select an element from matrices
myMatrix[1,2][1] 2
myMatrix[2,][1] 4 5 6
myMatrix[,1][1] 1 4
dim(myMatrix)[1] 2 3
myMatrix2 <- matrix(c(-6:-1), nrow = 2, ncol = 3, byrow = TRUE)
myMatrix + myMatrix2 [,1] [,2] [,3]
[1,] -5 -3 -1
[2,] 1 3 5
myMatrix %*% c(1,2,3) [,1]
[1,] 14
[2,] 32
diag(myMatrix2) #diagonal [1] -6 -2
t(myMatrix2) #transpose of matrix [,1] [,2]
[1,] -6 -3
[2,] -5 -2
[3,] -4 -1
- Arrays can be of any number of dimensions
myArray <- array(c('uno','dos', 'tres'), dim = c(3,3,3))
print(myArray), , 1
[,1] [,2] [,3]
[1,] "uno" "uno" "uno"
[2,] "dos" "dos" "dos"
[3,] "tres" "tres" "tres"
, , 2
[,1] [,2] [,3]
[1,] "uno" "uno" "uno"
[2,] "dos" "dos" "dos"
[3,] "tres" "tres" "tres"
, , 3
[,1] [,2] [,3]
[1,] "uno" "uno" "uno"
[2,] "dos" "dos" "dos"
[3,] "tres" "tres" "tres"
- They are categorical variables.
- You can create using a vector. Factors stores the vector along with distinct values of the elements in the vector as labesl.
- The labels are always character irrespective of wheter it is numeric or character or Boolean etc. input vector. They are useful in statistical modelling.
myVector <- c('blue', 'red', 'violet', 'red', 'red', 'blue')
print(myVector)[1] "blue" "red" "violet" "red" "red" "blue"
myVectorFactor <- factor(myVector)
print(myVectorFactor)[1] blue red violet red red blue
Levels: blue red violet
print(nlevels(myVectorFactor))[1] 3
Tidy Data
- Each variable forms a column
- Each observation forms a row
- Each type of observational unit forms a table
mySurvey <- data.frame(
name = c("Harry", "Ron", "Hermione", "Draco", "Cedric"),
gender = c("Male", "Male", "Female", "Male", "Male"),
age = c(11, 11, 11, 11, 12),
bloodStatus = c("Half-blood", "Pure-blood","Muggle-born", "Pure-blood", NA),
house = c("Gryffindor", "Gryffindor", "Gryffindor", "Slytherin", NA)
)
print(mySurvey) name gender age bloodStatus house
1 Harry Male 11 Half-blood Gryffindor
2 Ron Male 11 Pure-blood Gryffindor
3 Hermione Female 11 Muggle-born Gryffindor
4 Draco Male 11 Pure-blood Slytherin
5 Cedric Male 12 <NA> <NA>
is.na(mySurvey) name gender age bloodStatus house
[1,] FALSE FALSE FALSE FALSE FALSE
[2,] FALSE FALSE FALSE FALSE FALSE
[3,] FALSE FALSE FALSE FALSE FALSE
[4,] FALSE FALSE FALSE FALSE FALSE
[5,] FALSE FALSE FALSE TRUE TRUE
mySurvey[5,4] <- "Pure-blood"
str(mySurvey)'data.frame': 5 obs. of 5 variables:
$ name : Factor w/ 5 levels "Cedric","Draco",..: 3 5 4 2 1
$ gender : Factor w/ 2 levels "Female","Male": 2 2 1 2 2
$ age : num 11 11 11 11 12
$ bloodStatus: Factor w/ 3 levels "Half-blood","Muggle-born",..: 1 3 2 3 3
$ house : Factor w/ 2 levels "Gryffindor","Slytherin": 1 1 1 2 NA
levels(mySurvey$house) <- c("Gryffindor", "Slytherin", "Hufflepuff")
mySurvey[5,5] <- "Hufflepuff"
print(mySurvey) name gender age bloodStatus house
1 Harry Male 11 Half-blood Gryffindor
2 Ron Male 11 Pure-blood Gryffindor
3 Hermione Female 11 Muggle-born Gryffindor
4 Draco Male 11 Pure-blood Slytherin
5 Cedric Male 12 Pure-blood Hufflepuff
mySurvey$name[1] Harry Ron Hermione Draco Cedric
Levels: Cedric Draco Harry Hermione Ron
head(mySurvey) name gender age bloodStatus house
1 Harry Male 11 Half-blood Gryffindor
2 Ron Male 11 Pure-blood Gryffindor
3 Hermione Female 11 Muggle-born Gryffindor
4 Draco Male 11 Pure-blood Slytherin
5 Cedric Male 12 Pure-blood Hufflepuff
tail(mySurvey) name gender age bloodStatus house
1 Harry Male 11 Half-blood Gryffindor
2 Ron Male 11 Pure-blood Gryffindor
3 Hermione Female 11 Muggle-born Gryffindor
4 Draco Male 11 Pure-blood Slytherin
5 Cedric Male 12 Pure-blood Hufflepuff
- look for:
x %in% y, !(x %in% y), !x, !is.na()- R has a lot of packages for you to make some work easily done!
- CRAN is the name of the R package repository but you can find and download many R packages on Github.
- These packages are about: statistics, modelling, visualisation, manipulating, documentation, making websites/applications etc.
- dplyr, broom, tidyr, lubridate
- ggplot2
- purrr, magrittr, forecats, tibble
- readxl
install.packages('tidyverse')
library(tidyverse)-
R has many example datasets and you can look at the list of datasets from here to make practice.
-
Today we will use 20170930dataset and learn how to load a data set on working directory and manipulate it with dplyr functions.
- To read a dataset, you can use these functions, according to your file type: read.csv, read.table, read.xls, read.xlsx etc.
- You need path of the file: where you store your file. For example (/Users/hazelkavili/Desktop/R-LadiesIstanbul/20170930dataset.csv)
myDataset <- read.csv(file = "~/Desktop/R-LadiesIstanbul/20170930dataset.csv", sep = ",", header = TRUE)
print(myDataset) GlobalRank Company Sales Profits Assets MarketValue
1 307 KocHolding 47.1 1.3 61.1 13.6
2 333 IsBankasi 12.4 1.9 112.7 16.8
3 341 GarantiBankasi 9.8 1.9 99.2 22.0
4 382 Akbank 8.2 1.7 91.6 20.6
5 388 SabanciHolding 14.6 1.0 98.3 12.7
6 543 Halkbank 6.2 1.5 61.0 13.3
7 711 VakifBank 6.3 0.8 60.6 7.9
8 802 TurkTelekom 7.1 1.5 9.5 15.0
9 843 Turkcell 5.9 1.2 10.5 14.8
10 1210 TurkishAirlines 8.3 0.6 10.5 4.8
11 1245 Enka 5.7 0.6 8.2 8.3
12 1788 AnadoluEfes 3.6 0.3 6.5 9.0
13 1972 BIM 5.5 0.2 1.2 7.5
14 1977 FordOtosan 5.5 0.4 2.6 4.5
str(myDataset)'data.frame': 14 obs. of 6 variables:
$ GlobalRank : int 307 333 341 382 388 543 711 802 843 1210 ...
$ Company : Factor w/ 14 levels "Akbank","AnadoluEfes",..: 9 8 6 1 10 7 14 11 12 13 ...
$ Sales : num 47.1 12.4 9.8 8.2 14.6 6.2 6.3 7.1 5.9 8.3 ...
$ Profits : num 1.3 1.9 1.9 1.7 1 1.5 0.8 1.5 1.2 0.6 ...
$ Assets : num 61.1 112.7 99.2 91.6 98.3 ...
$ MarketValue: num 13.6 16.8 22 20.6 12.7 13.3 7.9 15 14.8 4.8 ...
summary(myDataset) GlobalRank Company Sales Profits
Min. : 307.0 Akbank :1 Min. : 3.600 Min. :0.200
1st Qu.: 383.5 AnadoluEfes :1 1st Qu.: 5.750 1st Qu.:0.600
Median : 756.5 BIM :1 Median : 6.700 Median :1.100
Mean : 917.3 Enka :1 Mean :10.443 Mean :1.064
3rd Qu.:1236.2 FordOtosan :1 3rd Qu.: 9.425 3rd Qu.:1.500
Max. :1977.0 GarantiBankasi:1 Max. :47.100 Max. :1.900
(Other) :8
Assets MarketValue
Min. : 1.200 Min. : 4.50
1st Qu.: 8.525 1st Qu.: 8.00
Median : 35.550 Median :13.00
Mean : 45.250 Mean :12.20
3rd Qu.: 83.975 3rd Qu.:14.95
Max. :112.700 Max. :22.00
-
Variables(columns)
- select
- mutate
-
Observations (rows)
- filter
- arrange
-
Groups
- summarise
look for Hadley's book for more magical words
- Basically tells R to take the value of that which is to the left and pass it to the right as an argument.
- cmd + shft + m
- kntr + shft + m
library(dplyr)
myDataset %>% filter(MarketValue > 5) %>% summarise(Average = mean(Assets)) Average
1 51.7
- Choosing is not losing!
select(dataframe, var1, var2, ...)
select(dataframe, 1:4, -2)smallSet <- myDataset %>% select(Company, MarketValue)
print(smallSet) Company MarketValue
1 KocHolding 13.6
2 IsBankasi 16.8
3 GarantiBankasi 22.0
4 Akbank 20.6
5 SabanciHolding 12.7
6 Halkbank 13.3
7 VakifBank 7.9
8 TurkTelekom 15.0
9 Turkcell 14.8
10 TurkishAirlines 4.8
11 Enka 8.3
12 AnadoluEfes 9.0
13 BIM 7.5
14 FordOtosan 4.5
- There are also helper functions: starts_with, end_with, contains
- Deals with info in your data which is not display
mutate(dataframe, newVariable = var1 + var2)
mutate(dataframe, x = a + b, y = x + c)mutateSet <- myDataset %>% mutate(TotalMoney = Assets + Profits)
head(mutateSet) GlobalRank Company Sales Profits Assets MarketValue TotalMoney
1 307 KocHolding 47.1 1.3 61.1 13.6 62.4
2 333 IsBankasi 12.4 1.9 112.7 16.8 114.6
3 341 GarantiBankasi 9.8 1.9 99.2 22.0 101.1
4 382 Akbank 8.2 1.7 91.6 20.6 93.3
5 388 SabanciHolding 14.6 1.0 98.3 12.7 99.3
6 543 Halkbank 6.2 1.5 61.0 13.3 62.5
- Filter out rows, specific type of observation
filter(dataframe, logicaltest)filteredSet <- myDataset %>% filter(Assets < 10 & Profits > 1)
print(filteredSet) GlobalRank Company Sales Profits Assets MarketValue
1 802 TurkTelekom 7.1 1.5 9.5 15
- Helps order observation (default ascending)
arrange(dataframe, var1)
arrange(dataframe, var1, desc(var2))arrangedSet <- myDataset %>% select(GlobalRank, Company, MarketValue) %>% arrange(desc(MarketValue))
print(arrangedSet) GlobalRank Company MarketValue
1 341 GarantiBankasi 22.0
2 382 Akbank 20.6
3 333 IsBankasi 16.8
4 802 TurkTelekom 15.0
5 843 Turkcell 14.8
6 307 KocHolding 13.6
7 543 Halkbank 13.3
8 388 SabanciHolding 12.7
9 1788 AnadoluEfes 9.0
10 1245 Enka 8.3
11 711 VakifBank 7.9
12 1972 BIM 7.5
13 1210 TurkishAirlines 4.8
14 1977 FordOtosan 4.5
- Helps order observation (default ascending)
summarise(dataframe, newVar = expression,. . .)
summarise(dataframe, sum = sum(A), avg = mean(B)..)summarisedSet <- myDataset %>% filter(grepl('Bank', Company) | grepl('bank', Company)) %>%
summarise(averageAssets = mean(Assets), averageSales = mean(Sales))
print(summarisedSet) averageAssets averageSales
1 85.02 8.58
look for grepl
- Mine Cetinkaya-Rundell's Teaching Data Science to New Users presentation
- Ismail Sezen's github
- data.world for many datasets
- Google :)
- Data Carpentary
========================================================