install-libraries.Rmd

---
title: "Installation of Data Science"
author: Emmanuel-R8
---

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

# Intro

Following are installation steps to locally 'compile' this bookdown. The instructions are targeted for linux users. Instructions will be added for Windows and Mac users at a later stage.

Each installation of R and RStudio are different, so it is likely that those steps will fail for some of them.

# Installation of the required packages

## Environment

R and RStudio are expected to be installed. An internet connection is required ;-). 

OPTIONAL: To render this installation file, the `rmarkdown` package is required, but the instructions can be read on a normal text editor. 

On a linux system, some packages/tools are required:

- Normal development environment, e.g. what the `build-essential` package provides on Debian/Ubuntu.

- curl, git, [...]

- development headers for libpoppler-cpp-dev

Make sure that the current working directory is the top of the tree, that is, just under `dsbook`. 

```{r}
# This is a small utility function to avoid re-installing packages
install_pkg <- function(name, ...) {
  if (!requireNamespace(name, quietly = TRUE)){
    install.packages(name, ...)
  }
}
```

Now that `packrat` is installed and initialised, `devtools` can be re-installed within the `packrat` environment.

The instructions list only a few packages, but they will in turn pull many others. During installation, a number of packages will require some development headers from your Linux distribution. 

## Installation of the relevant packages

The tidyverse packages is the central platform the book builds on.

```{r}
install_pkg("tidyverse")
install_pkg("tidytext")
```

Machine learning packages:

```{r}
install_pkg("caret")
install_pkg("lpSolve")
install_pkg("randomForest")
install_pkg("gam")
```

Additional visualisation packages and `ggplot2` addins:

```{r}
install_pkg("scatterplot3d")
install_pkg("VennDiagram")
install_pkg("maps")
install_pkg("mapproj")
install_pkg("animation")

install_pkg("gridExtra")
install_pkg("ggthemes")
install_pkg("ggrepel")
install_pkg("ggridges")
install_pkg("gganimate")
devtools::install_github("rensa/ggflags")
```

Some various utilities:

```{r}
install_pkg("devtools")
install_pkg("kableExtra")
install_pkg("matrixStats")
install_pkg("gtools")
install_pkg("pdftools")
install_pkg("english")
```

Now the `dslab` package that provides the information and scripts presented in the book can be downloaded.

```{r}
install_pkg("dslabs")
install_pkg("rafalib")
```

A package from the BioConductor repository:

```{r}
install_pkg("BiocManager")
BiocManager::install("genefilter")
```

A number of datasets are used throughout the book.

```{r}
install_pkg("gutenbergr")
install_pkg("Lahman")
install_pkg("countrycode")
install_pkg("NHANES")
install_pkg("HistData")
install_pkg("textdata")
```

Finally, the `tidytext` text mining package requires external dictionaries for sentiment analysis. Those require agreeing to licenses.

```{r}
tidytext::get_sentiments(lexicon = "afinn")
tidytext::get_sentiments(lexicon = "bing")
tidytext::get_sentiments(lexicon = "nrc")
```

## Book creation packages

`bookdown` generates the final output. `tufte` provides a style that Edward Tufte uses in his books and handouts. It is not the default style.

```{r}
install_pkg("bookdown")
install_pkg("tufte")
```

# Book generation

**The whole generation process is sometimes hit and miss. Re-running the generation commands sometimes solves prolems.**

The following is to generate a pdf version. The format can be customised by changing options at the top of the `index.Rmd` file.

```{r}
bookdown::render_book("index.Rmd", output_format = "bookdown::pdf_book")
```

If the first build succeeds, you could in principle to speed up the next generation, add `clean_envir = FALSE` to the options passed to `render_book`. 

Likewise, this generates a Gitbook...

```{r}
bookdown::render_book("index.Rmd", output_format = "bookdown::gitbook", clean_envir=FALSE)
```

... or an e-book.

```{r}
bookdown::render_book("index.Rmd", output_format = "bookdown::epub_book", clean_envir=FALSE)
```

```{r}
# Not working yet
# bookdown::render_book("index.Rmd", output_format="tufte::tufte_book", clean_envir=FALSE)
```