MovieLens.rmd

--- 
title: "HarvardX - PH125.9x Data Science: Capstone - Movie Lens"
subtitle: "HarvardX - PH125.9x Data Science Capstone"
author: "Emmanuel Rialland - https://github.com/Emmanuel_R8"
date: "`r format(Sys.time(), '%B %d, %Y')`"
site: bookdown::bookdown_site
output: 
  bookdown::pdf_book:
    toc: true
    toc_depth: 3
    number_sections: true
    keep_tex: true
    df_print: kable
  html_document: 
    number_sections: true
    df_print: paged
    theme: united
    
documentclass: report
geometry: "a4paper,left=2.5cm,right=2.5cm,top=2.5cm,bottom=2.5cm"
lot: yes
lof: yes

fontsize: 11pt
mainfont: "Lato"
monofont: "Hack"
monofontoptions: "Scale=0.7"

bibliography: [packages.bib, bibliography.bib]
biblio-style: apalike
link-citations: yes
colorlinks: yes

github-repo: Emmanuel_R8/HarvardX-Movielens
description: "HarvardX - PH125.9x Data Science Capstone"
---

```{r,echo=FALSE,message=FALSE}
###################################################################################################
## GLOBAL SETUP
## 

# Trigger line numbering
knitr::opts_chunk$set(
  class.source = "numberLines lineAnchors", 
  class.output = c("numberLines lineAnchors chunkout") 
)

# knitr global settings - By default, the final document will not include source code unless
# expressly stated.
knitr::opts_chunk$set(
  # Chunks
  eval = TRUE,
  cache = TRUE,
  echo = FALSE,
  message = FALSE,
  warning = FALSE,

  # filepaths
  fig.path =   'build/figure/graphics-', 
  cache.path = 'build/cache/graphics-', 
  
  # Graphics
  out.width = "70%",
  fig.align = "center",
  # fig.height = 3,

  # Text size
  size = "small"
)

if (knitr::is_html_output()) {
  knitr::opts_chunk$set(dev = "png")
} else {
  knitr::opts_chunk$set(dev = "pdf")
}

# Modify the size of the code chunks
# https://stackoverflow.com/questions/25646333/code-chunk-font-size-in-rmarkdown-with-knitr-and-latex
def.chunk.hook <- knitr::knit_hooks$get("chunk")

knitr::knit_hooks$set(chunk = function(x, options) {
  x <- def.chunk.hook(x, options)
  ifelse(options$size != "normalsize", paste0("\n \\", options$size, "\n\n", x, "\n\n \\normalsize"), x)
})

```


```{css echo=FALSE}
code {
  font-family: Hack, monospace;
  font-size: 85%;
  padding: 0;
  padding-top: 0.2em;
  padding-bottom: 0.2em;
  background-color: rgba(0,0,0,0.04);
  border-radius: 3px;
}

code:before,
code:after {
  letter-spacing: -0.2em;
  content: "\00a0";
}

```

<!--- Better HTML output --->
```{css echo=FALSE}
pre {
  font-family: Hack, monospace;
  font-size: 75%;
  margin-left: 15%; 
  #padding: 25%;
  padding-top: 0.2em;
  padding-bottom: 0.2em;
  background-color: rgba(0,0,255,0.1);
  #border-radius: 10px;
}

```



```{r preamble,echo=FALSE,message=FALSE}
### Preamble

# Usual libraries
library(tidyverse)
library(lubridate) 
library(gridExtra)

# Evaluation metrics including RMSE
library(Metrics)

# Data partitioning, machine learning models, training and prediction
library(caret)

# Use ds_theme_set()
library(dslabs)
ds_theme_set()

# Make things pretty
library(kableExtra)
library(RColorBrewer)
library(corrplot)

```


```{r echo=FALSE,message=FALSE}
## Coding style
# - The code makes heavy uses of tidyverse functions even when there would be more concise 
#   ways to write it. As the expression goes _"Premature optimisation is the root of all evil."_
#   
# - local({ ... }) is used when temporary variables are needed. Executing within a local
# scope avoids cluttering the environment with those variables. 
```