20170930_RForbeginners.Rpres

R-Ladies Istanbul: R For Beginners
========================================================
author: Hazel Kavili
date: 2017-09-30
width: 1400
height: 1280
font-family: 'Helvetica'

```{r, eval = TRUE, echo=FALSE, results=FALSE}
library(tidyverse)
```

R and R Studio
========================================================

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!

Looking for Help!
========================================================
- Ask Google
- Search in [Stackoverflow](https://stackoverflow.com/questions/tagged/r)
- [An introduction to R](http://cran.us.r-project.org/doc/manuals/R-intro.pdf)
- [R for Data Science](http://r4ds.had.co.nz) book from Hadley Wickham & Garrett Grolemund
- [Try R](http://tryr.codeschool.com)
- R mailing list: first, learn how to ask questions!
- [R Tutorials](http://www.tutorialspoint.com/r/)
- Get help from R:

```{r, eval = FALSE}
help.start()
help(mean)
?mean
example(mean)
```

Getting Started - I
========================================================
### **Motor Trend Car Road Tests - simple dataset**
```{r}
head(mtcars)
```
```{r}
tail(mtcars)
```

Getting Started - II
========================================================
### **Teaching base or tidyverse way**

Base R codes:
```{r, eval = FALSE}
mtcars$transmission <- 
  ifelse(mtcars$am == 0, "automatic", "manual")
```

dplyr codes:
```{r, eval = FALSE}
mtcars <- mtcars %>%
  mutate(transmission = case_when(am == 0 ~ "automatic", am == 1 ~ "manual"))
```

Getting Started - III
========================================================
### **Teaching base or tidyverse way**
#### **Visualisation**
Base R codes:
```{r, eval = FALSE}
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:
```{r, eval = FALSE}
p1 <- ggplot(mtcars, aes(x = disp, y = mpg, color = transmission)) +
       geom_point()

ggsave("plots/scatter_tidy.pdf", p1, width = 5, height = 3)
```


Getting Started - IV
========================================================

**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

Getting Started - V
========================================================

**Assignment**  and **Basic Operators**
- use **<-** for assigments
- +,-, *, /, ^, %%

**Logical Operators**
- <,>, <=, >=, ==, !=, !x, x & y, x | y

**Others**
- sum, sqrt, min, max, mean, var, sd, abs, summary

Most Frequently Used Objects
========================================================
- Vectors
- List
- Matrices
- Arrays
- Factors
- Data Frames

Vectors
========================================================
- use **c()** for concatenate more than one element.
- in programming vectors are variable sized sequence of values (not necessarily numbers).
```{r}
books <- c("history", "sci-fi", "fantasy")
print(books)

print(class(books))
```
```{r}
ages <- c(12,13,14,15,9,8)
print(ages)
print(class(ages))
```



Examples
========================================================
```{r}
#this is R-Ladies Istanbul
X <- 10
x <- 5
print(paste("X is", X))
print(paste("x is", x))
cat("X and x are equal? = ", X == x)
```

```{r}
myNumbers <- c(1:10) # c is short for concatenate
rep(myNumbers, times = 3)
twice <- rep(myNumbers, each = 2)
print(twice)
```

Examples - II
========================================================
```{r}
y <- c(1,2,3,10,15,20)
z <- c(y,4,5,6,y)
print(y)
```

- Vector Arithmetic
```{r}
5/y
5*y
y^2
sqrt(y)
```

Examples - III
========================================================
```{r}
seq(from = -10, to = 10, by = 1)
seq(from = 1, length = 25, by = 2 )
```

Simple Manipulations
========================================================

```{r}
weights <- c(55, 60, 45, 70, 56, 73, 59, 82)
sum(weights)
mean(weights)
sd(weights)
var(weights)
length(weights)
```

Lists
========================================================
- Lists can contain many different types of elements inside.
```{r}
myList <- list(c(1,2,3), 15, "hello")
print(myList)
```

- Select an element from lists
```{r}
myList[1]
myList[[1]][2]
myList[2]
```


Matrices
========================================================
- A matrix is **two-dimensional** recteangular data set.
```{r}
myMatrix <- matrix(c(1,2,3,4,5,6), nrow = 2, ncol = 3, byrow = TRUE)
print(myMatrix)
```

- Select an element from matrices
```{r}
myMatrix[1,2]
myMatrix[2,]
myMatrix[,1]
dim(myMatrix)
```

Matrices - II
========================================================
```{r}
myMatrix2 <- matrix(c(-6:-1), nrow = 2, ncol = 3, byrow = TRUE)
myMatrix + myMatrix2
```
```{r}
myMatrix %*% c(1,2,3)
diag(myMatrix2) #diagonal    
t(myMatrix2)  #transpose of matrix
```


Arrays
========================================================
- Arrays can be of any number of dimensions
```{r}
myArray <- array(c('uno','dos', 'tres'), dim = c(3,3,3))
print(myArray)
```

Factors
========================================================
- 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.

```{r}
myVector <- c('blue', 'red', 'violet', 'red', 'red', 'blue')
print(myVector)
myVectorFactor <- factor(myVector)
print(myVectorFactor)
print(nlevels(myVectorFactor))
```

Tidy Data Concept and Data Frames
========================================================
**Tidy Data**
- Each variable forms a column
- Each observation forms a row
- Each type of observational unit forms a table

```{r}
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)
```

Indexing, Selecting and Modifying
========================================================

```{r}
is.na(mySurvey)
mySurvey[5,4] <- "Pure-blood"
str(mySurvey)
levels(mySurvey$house) <- c("Gryffindor", "Slytherin", "Hufflepuff")
mySurvey[5,5] <- "Hufflepuff"
print(mySurvey)
```

Indexing, Selecting and Modifying - II
========================================================
```{r}
mySurvey$name
head(mySurvey)
tail(mySurvey)
```
- look for:
```{r, eval = FALSE}
x %in% y, !(x %in% y), !x, !is.na()
```

R Packages
========================================================
- 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.

Tidyverse
========================================================
- dplyr, broom, tidyr, lubridate 
- ggplot2
- purrr, magrittr, forecats, tibble
- readxl

Installation 
========================================================

```{r, eval = FALSE}
install.packages('tidyverse')
library(tidyverse)
```


Datasets
========================================================
- R has many example datasets and you can look at the list of datasets from [here](https://stat.ethz.ch/R-manual/R-devel/library/datasets/html/00Index.html) 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)*


```{r}
myDataset <- read.csv(file = "~/Desktop/R-LadiesIstanbul/20170930dataset.csv", sep = ",", header = TRUE)
print(myDataset)
```


Structure of Dataset
========================================================
```{r}
str(myDataset)
summary(myDataset)
```

Magical Words from dplyr!
========================================================
- **Variables(columns)**
  - select
  - mutate

- **Observations (rows)**
  - filter
  - arrange

- **Groups**
  - summarise

*look for Hadley's book for more magical words*

 %>% (Pipe) Operator
========================================================
- 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

```{r}
library(dplyr)
myDataset %>% filter(MarketValue > 5) %>% summarise(Average = mean(Assets))
```

Select
========================================================
- Choosing is not losing! 
```{r, eval = FALSE, warning = FALSE}
select(dataframe, var1, var2, ...)
select(dataframe, 1:4, -2)
```

```{r}
smallSet <- myDataset %>% select(Company, MarketValue)
print(smallSet)
```

- There are also helper functions: **starts_with**, **end_with**, **contains**

Mutate
========================================================
- Deals with info in your data which is not display

```{r, eval = FALSE, warning = FALSE}
mutate(dataframe, newVariable = var1 + var2)
mutate(dataframe, x = a + b, y = x + c)
```

```{r}
mutateSet <- myDataset %>% mutate(TotalMoney = Assets + Profits)
head(mutateSet)
```

Filter
========================================================
- Filter out rows, specific type of observation
```{r, eval = FALSE, warning = FALSE}
filter(dataframe, logicaltest)
```

```{r}
filteredSet <- myDataset %>% filter(Assets < 10 & Profits > 1)
print(filteredSet)
```

Arrange
========================================================
- Helps order observation (default ascending)

```{r, eval = FALSE, warning = FALSE}
arrange(dataframe, var1)
arrange(dataframe, var1, desc(var2))
```

```{r}
arrangedSet <- myDataset %>% select(GlobalRank, Company, MarketValue) %>%  arrange(desc(MarketValue))
print(arrangedSet)
```

Summarise
========================================================
- Helps order observation (default ascending)

```{r, eval = FALSE, warning = FALSE}
summarise(dataframe, newVar = expression,. . .)
summarise(dataframe, sum = sum(A), avg = mean(B)..)
```

```{r}
summarisedSet <- myDataset %>% filter(grepl('Bank', Company) | grepl('bank', Company)) %>% 
  summarise(averageAssets = mean(Assets), averageSales = mean(Sales))

print(summarisedSet)
```

*look for grepl*


References
========================================================
- Mine Cetinkaya-Rundell's  [Teaching Data Science  to New Users](https://github.com/mine-cetinkaya-rundel/2017-07-05-teach-ds-to-new-user) presentation
- Ismail Sezen's [github](https://isezen.github.io/r-presentation/intro.html)
- [data.world](data.world) for many datasets
- Google :) 
- [Data Carpentary](http://www.datacarpentry.org/)


========================================================
# <center> **Any questions?** </center>

#### - hazel@rladies.org & istanbul@rladies.org
#### - twitter.com/RLadiesIstanbul
#### - meetup.com/rladies-istanbul