Session 3 - Functions and loops.Rmd

---
title: "Conditionals, Functions, and Loops"
author: "Ashwin Malshé"
date: "06/01/2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Conditionals

Like most other programming languages, R also has `if()` and `else()` statements. Let's look into a simple example. 

Generate a random number from a uniform random distribution. If the number is less than 0.5, print `Tails` otherwise print `Heads`

```{r}
a <- runif(1)
print (a)
if ( a < 0.5) {
  print("Tails")
  } else (print("Heads"))

```

R also has a vectorized `ifelse()`. This can speed up operations. For example, consider a vector A with 10 random numbers between 0 and 1. We want to create a character vector such that if a number in A is less than 0.5, we return "Tails", otherwise "Heads".

With `if` and `else`, we will have to evaluate this separately for each element of A. `ifelse` can do it in one shot.

```{r}
A <- runif(10)
A
B <- ifelse(A < 0.5, "Tails", "Heads")
B
```
*This looks like the formula in Excel!*

## Functions

Functions wrap longer procedures in one single command

Next we will create a function called forever21.

```{r}
forever21 <- function(age) {
  if (age <= 21) {
    return(age)
  } else {
    return(21)
  }
}
```


```{r}
forever21(50)
forever21(15)
```

Write a function to add two numbers

```{r}
mySum <- function(a, b) {
  return(a + b)
}

mySum(5, 10)
mySum(255, 255)
```

Write a function to concatenate two character variables with three delimiters.
For example, you input "Hello" and "world!" separately and you get "Hello world!" In this case the delimiter is space.


```{r}
myWord <- function(a, b) {
  w1 <- paste(a, b, sep = "|")
  w2 <- paste(a, b, sep = "-")
  w3 <- paste(a, b, sep = "***")
  return(c(w1, w2, w3))
}
myWord("Hello", "world!")
```

Write a function to winsorize a vector at a given percentage. Winsorization is a common data science technique to mitigate the effect of outliers. Rather than deleting these observations, we can set them to a a set percentile. For instance,  if we want to winsorize a variable at 90%, we will set all the values lower than bottom 5% to the 5th percentile and all the values higher than top 5% to the 95th percentile. Thus, we keep middle 90% unchanged.


```{r}
winsor <- function(vec, w = 0.95) {
  l1 <- (1 - w) / 2
  l2 <- 1 - l1
  q <- quantile(vec, c(l1, l2), type = 1)
  vec1 <- ifelse(vec < q[1], q[1], vec)
  vec2 <- ifelse(vec1 > q[2], q[2], vec1)
  return(vec2)
}
```



```{r}
winsor(vec = c(1:100), w = 0.9)
```



## Writing functions with default values

Write a function for area of a circle with radius r. The default value should be for radius r = 1

```{r}
cirArea <- function(r = 1) pi * r^2
```

get area of a circle with radius = 5. Verify the result

```{r}
# Answer using the function
cirArea(5)
# Answer using the raw formula
pi * 5^2

identical(cirArea(5), pi * 5^2)
```


## Saving and retrieving functions
Copy the function and paste it in an R script. Save the file on your hard disk and make note of the path to that file.

When you want to load the function, use 

`source("path name to the function R file")`


## Looping in R

```{r}
for (i in 1:10) {
  print(paste("My number is", i))
}
```

Also check out `while` command in R. It can be used when the condition is not easy or possible to write in a `for`. For example, you may want to keep running the loop until specific time (e.g., 5 pm) or until you get a specific number of observations from an operation (e.g., download 5,000 tweets)

```{r}
i <- 1
while(i <= 10) {
  print(paste("My number is", i))
  i <- i + 1
}
```