Loss modelling, reserving and fraud analytics in R

# Loss modelling, reserving and fraud analytics in R - Prework
## A hands-on workshop
<html>
<div style="float:left">

</div>
<hr align='center' color='#116E8A' size=1px width=97%>
</html>
### Katrien Antonio & Jonas Crevecoeur
### <a href="https://www.github.com/katrienantonio/workshop-loss-reserv-fraud">IA|BE workshop</a> | June 3, 10 & 17, 2021

---

# Prologue

---

# Introduction

### Course

The course repo on GitHub, where you can find the data sets, lecture sheets, R scripts and R markdown files.

### Us

---

# Checklist

☑ Do you have a fairly recent version of R?
  
  ```r
  version$version.string
  ## [1] "R version 4.0.3 (2020-10-10)"
  ```

☑ Do you have a fairly recent version of RStudio? 
  
  ```r
  RStudio.Version()$version
  ## Requires an interactive session but should return something like "[1] ‘1.3.1093’"
  ```

☑ Have you installed the R packages listed in the software requirements?

☑ Have you created an account on RStudio Cloud (to avoid any local installation issues)?
  
---

# What's out there - the R universe

---

# What is R?

> <font size="+2"> <p align="justify">The R environment is an integrated suite of software facilities for data manipulation, calculation and graphical display.</p></font>

</br>

A brief history:

- R is a dialect of the S language.

- R was written by .KULbginline[R]obert Gentleman and .KULbginline[R]oss Ihaka in 1992.

- The R source code was first released in 1995.

- In 1998, the Comprehensive R Archive Network [CRAN](http://CRAN.R-project.org/) was established.

- The first official release, R version 1.0.0, dates to 2000-02-29. Currently R 4.0.3 (October, 2020).

- R is open source via the [GNU General Public License](https://en.wikipedia.org/wiki/GNU_General_Public_License).

---

# Explore the R architecture

- R is like a car's engine

- RStudio is like a car's dashboard, an integrated development environment (IDE) for R.

R: Engine            |  RStudio: Dashboard
:-------------------------:|:-------------------------:
<img src="image/engine.jpg" alt="Drawing" style="height: 300px;"/>  |  <img src="image/dashboard.jpg" alt="Drawing" style="height: 300px;"/>

---

# How do I code in R?

Keep in mind:

- unlike other software like Excel, STATA, or SAS, R is an interpreted language

- no point and click in R!

- .KULbginline[you have to program in R]!

R .KULbginline[packages] extend the functionality of R by providing additional functions, and can be downloaded for free from the internet.

R: A new phone           |  R Packages: Apps you can download
:-------------------------:|:-------------------------:
<img src="image/iphone.jpg" alt="Drawing" style="height: 150px;"/>  |  <img src="image/apps.jpg" alt="Drawing" style="height: 150px;"/>

---

# How to install and load an R package?

Install the {ggplot2} package for data visualisation

```r
install.packages("ggplot2")
```

Load the installed package

```r
library(ggplot2)
```

And give it a try

```r
head(diamonds)
ggplot(diamonds, aes(clarity, fill = cut)) + 
  geom_bar() + theme_bw()
```

Packages are developed and maintained by R users worldwide.

They are shared with the R community through CRAN: now 16,460 packages online (on November 2, 2020)!

]

]

---

# Why R and RStudio?

### Data science positivism

- Next to Python, R has become the *de facto* language for data science, with a cutting edge *machine learning toolbox*.
- See: [The Popularity of Data Science Software](http://r4stats.com/articles/popularity/)
- R is open-source with a very active community of users spanning academia and industry.

### Bridge to actuarial science, econometrics and other tools

- R has all of the statistics and econometrics support, and is amazingly adaptable as a “glue” language to other programming languages and APIs.
- R does not try to be everything to everyone. The RStudio IDE and ecosystem allow for further, seemless integration (with e.g. python, keras, tensorflow or C).
- Widely used in actuarial undergraduate programs

### Disclaimer + Read more

- It's also the language that we know best.
- If you want to read more: [R-vs-Python](https://blog.rstudio.com/2019/12/17/r-vs-python-what-s-the-best-for-language-for-data-science/), [when to use Python or R](https://www.datacamp.com/community/blog/when-to-use-python-or-r) or [Hadley Wickham on the future of R](https://qz.com/1661487/hadley-wickham-on-the-future-of-r-python-and-the-tidyverse/)

---

background-image: url("image/tidyverse2.1.png")
background-size: cover
background-size: 65% 
background-position: center

---

# Welcome to the tidyverse!

><p align="justify">The .KULbginline[tidyverse] is an opinionated collection of R packages designed for data science. All packages share an underlying design philosophy, grammar, and data structures. </p>

More on: [tidyverse](https://www.tidyverse.org).

Install the packages with `install.packages("tidyverse")`. Then run `library(tidyverse)` to load the core tidyverse.

---

# Principles of tidy data

Three interrelated rules from the [R for data science](https://r4ds.had.co.nz/) book by Garrett Grolemund and Hadley Wickham:

1. Each variable must have its own column.
2. Each observation must have its own row.
3. Each value must have its own cell.

.footnote[This figure is taken from Chapter 12 on Tidy data in [R for data science](https://r4ds.had.co.nz/).]

---

# Workflow of a data scientist

Here is a model of the .hi-pink[tools needed in a typical data science project]:

> <p align="justify"> Together, tidying and transforming are called <b>wrangling</b>, because getting your data in a form that’s natural to work with often feels like a fight! </p>

> <p align="justify"> Models are complementary tools to visualisation. Once you have made your questions sufficiently precise, you can use a model to answer them. Models are a fundamentally <b>mathematical or computational tool</b>, so they generally scale well. But every model makes <b>assumptions</b>, and by its very nature a model cannot question its own assumptions. That means <b>a model cannot fundamentally surprise you</b>.</p>

---

# Data wrangling and visualisation

---

# A tibble instead of a data.frame <img src="image/tibble.png" class="title-hex">

Within the tidyverse `tibble` is a modern take on a `data.frame`:

- keep the features that have stood the test of time

- drop the features that used to be convenient but are now frustrating.

You can use:

- `tibble()` to create a new tibble

- `as_tibble()` transforms an object (e.g. a data frame) into a tibble.

Quick example: explore the differences!

```r
mtcars
# install.packages("tidyverse")
library(tidyverse)
as_tibble(mtcars)
```

---

# Chains with the pipe operator <img src="image/pipe.png" class="title-hex">

In R, the pipe operator is `%>%`.

It takes the output of one statement and makes it the input of the next statement.

When describing it, you can think of it as a “THEN”; with this operator it becomes easy to chain a sequence of calculations.

For example, when you have an input data and want to call functions `foo` and `bar` in sequence, you can write `data %>% foo %>% bar`.

A first example:

- take the `diamonds` data (from the {ggplot2} package)

- then subset

```r
diamonds %>% filter(cut == "Ideal")
```

Some excellent blog posts about this operator: [Pipes in R tutorial for beginners](https://www.datacamp.com/community/tutorials/pipe-r-tutorial) and [how to write this in base R](https://gist.github.com/hadley/c430501804349d382ce90754936ab8ec).

---

# Data manipulation verbs <img src="image/dplyr.png" class="title-hex">

The {dplyr} package holds many useful data manipulation verbs:

- `mutate()` adds new variables that are functions of existing variables

- `select()` picks variables based on their names

- `filter()` picks cases based on their values

- `summarise()` reduces multiple values down to a single summary

- `arrange()` changes the ordering of the rows.

These all combine naturally with `group_by()` which allows you to perform any operation “by group”.

A first example:

```r
diamonds %>% mutate(price_per_carat = price/carat) %>% filter(price_per_carat > 1500)
```

```r
diamonds %>% group_by(cut) %>% summarize(price = mean(price), carat = mean(carat))
```

---

<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/4.7.0/css/font-awesome.min.css">

## <i class="fa fa-edit"></i> <br> Your turn

]

To get warmed up, let's do some .KULbginline[basic explorations] of the {tidyverse} instructions. The idea is to get some feel for these functions.

1.1. Create a data frame (with `data.frame(.)`) or tibble (with `tibble(.)`) `df` with two variables `x` and `y`. Enter some values for these variables.

1.2. Create a new variable `z` that is the sum of `x` and `y`. Use `base` R instructions and then use the pipe operator and `mutate(.)`.

2.1. Create a new data vector `v` with some entries, use `c(.)`.

2.2. Try the following instructions:

```r
round(mean(x), 2)
mean(x) %>% round(2)
x %>% mean %>% round(2)
```

]

---

First, you put together the `data.frame`

```r
df <- data.frame(x = c(0, 1), y = c(0, 1))
df 
```

or the `tibble`

```r
df <- tibble(x = c(0, 1), y = c(0, 1))
df 
```

Next, you create a new variable

```r
df$z <- df$x + df$y 
df
```

or with `mutate(.)`

```r
df %>% mutate(z = x+y)
df
```

]

You create a vector `x` with some entries

```r
x <- c(0.157, 0.135, 0.359)
```

and then you evaluate

```r
round(mean(x), 2)
mean(x) %>% round(2)
x %>% mean %>% round(2)
```

These implementations all lead to the same result:

```
## [1] 0.22
```

Which one do you find most intuitive?

]

---

# Plots with ggplot2 <img src="image/ggplot2.png" class="title-hex">

The aim of the {ggplot2} package is to create elegant data visualisations using the .hi-pink[grammar of graphics].

Here are the basic steps:

- begin a plot with the function `ggplot()` creating a coordinate system that you can add layers to

- the first argument of `ggplot()` is the dataset to use in the graph

A first example

```r
library(ggplot2)
ggplot(data = mpg)
ggplot(mpg)
```

creates an empty graph.

You will now add layers to this graph!

---

# Plots with ggplot2 <img src="image/ggplot2.png" class="title-hex">

You complete your graph by adding one or more .hi-pink[layers] to `ggplot()`.

For example:

- `geom_point()` adds a layer of points to your plot, which creates a scatterplot

- `geom_smooth()` adds a smooth line

- `geom_bar` a bar plot

and many more, see [ggplot2 documentation](https://ggplot2.tidyverse.org/).

Each `geom` function in `ggplot2` takes an aesthetic mapping argument:

- maps variables in your dataset to visual properties

- always paired with `aes()` and the `$x$` and `$y$` arguments of `aes()` specify which variables to map to the `$x$` and `$y$` axes.

---

```r
library(ggplot2)
ggplot(mpg, `aes(displ, hwy, colour = class)`) + 
  `geom_point()` + `theme_bw()`

```

]

.pull-right[
  
Extend the empty graph now with (here: global) aesthetic mapping argument `aes(displ, hwy, colour = class)`.

This implies: `displ` on the x-axis, `hwy` on the y-axis and `class` to differentiate the color of the plotting symbol.

With `geom_point` you add a layer of points to the empty graph.

`theme_bw()` changes the `ggtheme` to a simple black-and-white theme.

]

---

# What else is there?

Recall

There are .KULbginline[(multiple) alternative ways] to do what the packages and functions in the tidyverse do.

For instance:

- base R
- the {data.table} package

You can read more about comparisons on e.g. [how to write this in base R](https://gist.github.com/hadley/c430501804349d382ce90754936ab8ec) or [Base R, the tidyverse, and data.table: a comparison of R dialects to wrangle your data](https://wetlandscapes.com/blog/a-comparison-of-r-dialects/).

---

# Thanks!  <img src="image/xaringan.png" class="title-hex">

Slides created with the R package [xaringan](https://github.com/yihui/xaringan).
<br> <br> <br>
Course material available via 
<br>
<svg style="height:0.8em;top:.04em;position:relative;fill:#116E8A;" viewBox="0 0 496 512"><path d="M165.9 397.4c0 2-2.3 3.6-5.2 3.6-3.3.3-5.6-1.3-5.6-3.6 0-2 2.3-3.6 5.2-3.6 3-.3 5.6 1.3 5.6 3.6zm-31.1-4.5c-.7 2 1.3 4.3 4.3 4.9 2.6 1 5.6 0 6.2-2s-1.3-4.3-4.3-5.2c-2.6-.7-5.5.3-6.2 2.3zm44.2-1.7c-2.9.7-4.9 2.6-4.6 4.9.3 2 2.9 3.3 5.9 2.6 2.9-.7 4.9-2.6 4.6-4.6-.3-1.9-3-3.2-5.9-2.9zM244.8 8C106.1 8 0 113.3 0 252c0 110.9 69.8 205.8 169.5 239.2 12.8 2.3 17.3-5.6 17.3-12.1 0-6.2-.3-40.4-.3-61.4 0 0-70 15-84.7-29.8 0 0-11.4-29.1-27.8-36.6 0 0-22.9-15.7 1.6-15.4 0 0 24.9 2 38.6 25.8 21.9 38.6 58.6 27.5 72.9 20.9 2.3-16 8.8-27.1 16-33.7-55.9-6.2-112.3-14.3-112.3-110.5 0-27.5 7.6-41.3 23.6-58.9-2.6-6.5-11.1-33.3 2.6-67.9 20.9-6.5 69 27 69 27 20-5.6 41.5-8.5 62.8-8.5s42.8 2.9 62.8 8.5c0 0 48.1-33.6 69-27 13.7 34.7 5.2 61.4 2.6 67.9 16 17.7 25.8 31.5 25.8 58.9 0 96.5-58.9 104.2-114.8 110.5 9.2 7.9 17 22.9 17 46.4 0 33.7-.3 75.4-.3 83.6 0 6.5 4.6 14.4 17.3 12.1C428.2 457.8 496 362.9 496 252 496 113.3 383.5 8 244.8 8zM97.2 352.9c-1.3 1-1 3.3.7 5.2 1.6 1.6 3.9 2.3 5.2 1 1.3-1 1-3.3-.7-5.2-1.6-1.6-3.9-2.3-5.2-1zm-10.8-8.1c-.7 1.3.3 2.9 2.3 3.9 1.6 1 3.6.7 4.3-.7.7-1.3-.3-2.9-2.3-3.9-2-.6-3.6-.3-4.3.7zm32.4 35.6c-1.6 1.3-1 4.3 1.3 6.2 2.3 2.3 5.2 2.6 6.5 1 1.3-1.3.7-4.3-1.3-6.2-2.2-2.3-5.2-2.6-6.5-1zm-11.4-14.7c-1.6 1-1.6 3.6 0 5.9 1.6 2.3 4.3 3.3 5.6 2.3 1.6-1.3 1.6-3.9 0-6.2-1.4-2.3-4-3.3-5.6-2z"/></svg> https://github.com/katrienantonio/workshop-loss-reserv-fraud-2020