Tidyverse and ggplot

class: center, middle, inverse, title-slide

# Tidyverse and ggplot
### Niels Richard Hansen
### February 6, 2020

---

layout: true

background-image: url(KUlogo.pdf)
background-size: cover
background-position: left

---

## Tidyverse packages

```r
library("tidyverse")
```

```
── Attaching packages ──────────────────────────────────────────────── tidyverse 1.3.0 ──
```

```
✓ ggplot2 3.2.1     ✓ purrr   0.3.3
✓ tibble  2.1.3     ✓ dplyr   0.8.3
✓ tidyr   1.0.2     ✓ stringr 1.4.0
✓ readr   1.3.1     ✓ forcats 0.4.0
```

```
── Conflicts ─────────────────────────────────────────────────── tidyverse_conflicts() ──
x dplyr::filter() masks stats::filter()
x dplyr::lag()    masks stats::lag()
```

The tidyverse package loads a number of R packages that have been developed
to support a cleaner, more efficient and more coherent framework for data handling and 
data transformation in R.

---

## Core tidyverse packages

<div id="htmlwidget-c2a0aad5fb3af0162398" style="width:100%;height:auto;" class="datatables html-widget"></div>
<script type="application/json" data-for="htmlwidget-c2a0aad5fb3af0162398">{"x":{"filter":"none","data":[["tibble","dplyr","tidyr","readr","purrr","ggplot2"],["A data structure for tabular data","Rearrange, filter, transform and join data tibbles","Gather or spread tables to make them tidy","Read and write tabular data","Functional programming","Graphics"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th>Package<\/th>\n      <th>Functionality<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"ordering":false,"searching":false,"lengthChange":false,"info":false,"paging":false,"columnDefs":[{"className":"dt-head-left","targets":"_all"}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

---

## Reading data

```r
## Using the readr package
traffic <- read_csv("./Data/traffic_small.csv")
```

```
Parsed with column specification:
cols(
  `machine name` = col_character(),
  `user id` = col_double(),
  size = col_double(),
  time = col_double(),
  date = col_date(format = ""),
  month = col_character()
)
```

This data set is from a [classical 1995 data set on Boston University WWW data transfers](Data/traffic_small.csv).

The `size` column contains sizes of downloaded documents in bytes 
and the `time` column contains the time in seconds.

---

## Data

```r
## This is a tibble
traffic 
```

```
# A tibble: 16,414 x 6
  `machine name` `user id`  size  time date       month  
  <chr>              <dbl> <dbl> <dbl> <date>     <chr>  
1 cs18              146579  2464 0.493 1995-04-24 1995-04
2 cs18              995988  7745 0.326 1995-04-24 1995-04
3 cs18              317649  6727 0.314 1995-04-24 1995-04
4 cs18              748501 13049 0.583 1995-04-24 1995-04
5 cs18              955815   356 0.259 1995-04-24 1995-04
6 cs18              596819 15063 0.336 1995-04-24 1995-04
# … with 1.641e+04 more rows
```

A tibble is a data table and behaves in many ways effectively as a data frame

---

## Exercise

Download the data set `ballon.txt` from Absalon.

Read the data set into R as

* a data frame using a base R function 
* a tibble using the readr package

In both cases, use the RStudio "Import Dataset" functionality to construct the 
code you need to load the data set.

After solving this exercise, you should have a data frame or a tibble in the 
workspace called `ballon`.

---

## ggplot

If you want to use ggplot, you need to load the package ggplot2 (or tidyverse).

```r
library(ggplot2)
```
--

You can make simple scatterplots quickly using `qplot` (quick plot).

```r
qplot(rnorm(100), rnorm(100))  ## Uhhh, grey theme :-(
```

---

## ggplot

I always use a different theme.

```r
theme_set(theme_bw())
```
--

Newbie ggplot users spotted by their use of the default grey theme!

```r
qplot(rnorm(100), rnorm(100))  ## Ahh, black-and-white theme :-)
```

---
	
## A simple bar chart

```r
p <- ggplot(data = traffic, aes(x = `machine name`)) + 
  geom_bar()  ## The `geom_bar` does tabulation and construct the 'count' variable
p
```

---
## Flipping the bar chart

```r
p + coord_flip()
```

---

## A simple histogram

```r
ggplot(data = traffic, aes(x = log10(time))) + 
  geom_histogram()
```

```
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
```

---

## Some manual control

```r
ggplot(data = traffic, aes(x = time)) + 
  geom_histogram(bins = 20) + 
  scale_x_log10()
```

---

## A density plot

```r
ggplot(data = traffic, aes(x = time, fill = `machine name`)) +
  geom_density() + 
  scale_x_log10()
```

---

## Controlling fill color and transparency

```r
ggplot(data = traffic, aes(x = time, fill = `machine name`)) +
  geom_density(alpha = 0.5) + 
  scale_x_log10() +
  scale_fill_manual(values = c("red", "blue", "yellow"))
```

---

## Stacking histograms

```r
ggplot(data = traffic, aes(x = time, fill = `machine name`)) +
  geom_density(position = "stack") + 
  scale_x_log10() +
  scale_fill_manual(values = c("red", "blue", "yellow"))
```

---

## Exercise

Construct a histogram and/or a density plot of the variable `Tid` in the 
ballon data set.

Use different colors in the plot for the different colors of balloons.

Experiment with stacking and/or transparency.

Can you match the color used in your plot to the color of the balloon?