---
title: "R for Data Science: Chapters 2.1 - 2.2"
subtitle: "Data 309" 
date: today
format: html
toc: true
toc-title: Sections 2.1 - 2.2
---

## RStudio
[Cheatsheet](https://ucdavis-bioinformatics-training.github.io/Oct2017-ILRI-Workshop/Cheat_Sheets/rstudio-IDE-cheatsheet.pdf)


## Make a script (or R - notebook) for each class

Shortcut: `Ctrl/Cmd - shift - N` or
point-and-click as shown 

![file -> open -> script](open.png)
Use comments to organize your work!

## Data Science Workflow

What are the main tools of data science?  One answer to this question is the following diagram.

![Figure: The data science flow](https://r4ds.hadley.nz/diagrams/data-science/whole-game.png "ds flow")


## Dataset: Gapminder

Let's explore a new dataset called <i>gapminder</i>. It concerns world development and is found at (<https://www.gapminder.org/>) 

**Gapminder**: *Fight devastating ignorance with a fact-based worldview everyone can understand.* 

```{r}
#| eval: FALSE
install.packages("gapminder")
```

```{r, message=FALSE, warning=FALSE}
library(gapminder)
library(tidyverse)
```

Below we look at a transposed version of the data with `glimpse`, which lives in the **dplyr** package, included in the tidyverse. Or `View()`, or or just type `gapminder`.

```{r}
glimpse(gapminder)
```

We see that this dataset has variables: countries, continents, year, life expectancy, population and GDP per capita. 

## Useful exploratory tips

1. Access variables directly `gapminder$country` or attach the data to our workspace to avoid the harsh syntax with `attach(gapminder)`. 
2. Then we can see all the countries listed with `unique(country)`.  
3. Recall, `?gapminder` is helpful to learn about the dataset.  
4. `names(gapminder)` lists all the variable names of gapminder

## Grammar of Graphics / ggplot syntax

One of the main ways to visualize data in R is **ggplot2**, which utilizes the conceptual framework of the **grammar of graphics**.  In English, grammar dictates that each sentence must have a subject and verb.  In the grammar of graphics, each plotting element must have **data**, **aesthetics** and a **geometry**.  Good sentences often have prepositions, adverbs, etc., and good graphics have more layers as well. 


![Figure: ggplot2 layers](ggplot2_layers.png)

Our first input to ggplot is always a data table.  Notice below, there is nothing to see, but also no error.  It's just a blank plot. 

```{r}
ggplot(data = gapminder)
```

The second input is a mapping argument called `aes()` or aesthetic.  (Another blank plot.)

```{r}
ggplot(
  data = gapminder,
  aes(x = year)
)
```

Only with the third layer do we get something interesting. 

```{r}
ggplot(
  data = gapminder,
  aes(x = year, y = lifeExp)) + 
  geom_point() 

```

This is a pretty bad plot.  We learn something, but we could learn the same with less.  Scatterplots are best used to compare **numerical data**, not categorical.  The variable GDP (gross domestic product) is numerical, so let's see if life expectancy is possibly related to GDP per capita.

```{r}
ggplot(
  data = gapminder,
  aes(x = gdpPercap, y = lifeExp)) + 
  geom_point() 

```

In the above plot, can you determine a relationship between GDP and Life Expectancy? Let's see if time plays a role.

```{r}
ggplot(
  data = gapminder,
  aes(x = gdpPercap, y = lifeExp, color = year)) + 
  geom_point() 

```

In the above, we included a 3rd variable as color.  Now apply a fourth layer, a statistic.  In this case a curve of best fit.  

```{r}
ggplot(
  data = gapminder,
  aes(x = gdpPercap, y = lifeExp)) + 
  geom_point(aes(color = year)) +
  geom_smooth()
```

We improve the graphic by adding labels such as a title, subtitle and more with `labs()`.  Do `?labs()` to explore possibilities.  How could the plot below be improved even further?

```{r}
ggplot(
  data = gapminder,
  aes(x = gdpPercap, y = lifeExp)) + 
  geom_point(aes(color = year)) +
  labs(title = "GDP vs. Life Expectancy", 
       y = "GDP per capita", 
       subtitle = "From 1960 to 2000")
```

## Assignment 2 

(i) Edit the file `lecture_1.qmd` in the [week1](http://euclid.nmu.edu/~joshthom/teaching/dat309/week1/) folder to contain responses to the following: (note you need to "render" the .qmd file into an .html file)

(ii) Recreate the final plot above but with a different mapping, such as GDP as a function of year and color by continent.  

(iii) What happens when you try to color by country?  

(iv) What makes some mappings more useful than others?