README.Rmd

---
title: "README.md"
author: "Daniel Chen"
date: ""
output:
  md_document:
    variant: markdown_github 
    toc: true
---

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

# Working with dataframes

```{r}

# load the ggplot2 library to get the diamonds dataset
library(ggplot2)

# look at the diamonds dataset
diamonds
```


You can see here that it prints out a portion of the dataset,
this is because this is an `tibble` object, which is an extension of the core
`data.frame` object in R.

If you work with any package that Hadley Whickham has written,
be prepared to see tibbles.

They (for the most part) will work exactly the same as `data.frame` objects


## Subset columns
```{r, results='hide'}
# get 1 column from the dataframe
diamonds$carat
```

I'm not showing the output for the code becuse there are too many values to print to the screen
The results of using the `$` notation for subsetting will be a `vector` of values

```{r}
# get 1 column using bracket notation
# bracket noation [row, columns]
diamonds[, 'carat']
```

You can visually see the difference between the `[ ]` and `$` notation of subsetting.
`[ ]` will return a dataframe (or in this case, a tibble) back

```{r}
# get multiple columns from the dataset
diamonds[, c('carat', 'price')]
```

## Subset rows
```{r}
# various ways of getting rows from data

diamonds[1, ]
diamonds[10, ]
diamonds[c(1, 2), ]
diamonds[c(11:20), ]
diamonds[11:20, ]
```

## Subset rows and columns

```{r}
# getting rows and columns from your data
diamonds[11:20, c('cut', 'color')]
```

## Getting a glimpse of your data

Tibbles are nice in that they don't print all the rows out to the screen by default.
If you end up working with other dataframe-like objects, they might print out more rows of data you will care to look at

```{r}
# get the first 6 rows of the dataset
head(diamonds)

# get the last 6 rows of the dataset
tail(diamonds)
```

```{r}
# get numner of rows and columns
dim(diamonds)

nrow(diamonds)
ncol(diamonds)

# get the number of columns from dim
dim(diamonds)[2]
```

# cars dataset
```{r}
# look at the built in r dataset, cars
head(cars)
```

```{r}
# use the class function to see what you have
class(diamonds)
class(cars)
```

# mtcars datset

```{r}
# the mtcars datset
head(mtcars)
```

```{r}
# get a specific value out of the data
mtcars[1, 'wt']
```

## getting just the columns names and row names
```{r}
names(mtcars)
row.names(mtcars)
```

```{r}
# specifying row and columns by name
mtcars["Mazda RX4", 'wt']
```

## Creating new columns
```{r}
# create new column called cars
mtcars$cars <- row.names(mtcars)

mtcars$cyl_wt <- mtcars$cyl / mtcars$wt

```

## Getting summary statistics on your data
summary(mtcars)

## Using `attach`

The reason why you would *not* want to use `attach` in your code is becuase it mucks up the variables
you have access to in your R code.

It can be very easy to forget which what the column names are in your dataset.

```{r}
# dont do this
attach(mtcars)
mpg
detach(mtcars)

# do this idead, if you want
with(mtcars, mpg)
```

# Conditonal Subsetting

We can perform conditional subsetting by using the way we subset dataframes: `[ , ]`

```{r}
summary(mtcars)
```

Let's subset our data and look for all the cars where the `mpg` is less than `15.43`

```{r}
mtcars[mtcars$mpg < 15.43, ]
```

The reason why this works is becuase the row subsetter returns a vector of `TRUE`/`FALSE` values

```{r}
mtcars$mpg < 15.43
```

When you give a dataframe row or column slicer a vector of booleans (i.e., TRUE/FALSE),
it will return the rows and/or columns that have a TRUE.

## Subsetting a single column with `[ ]`

Subsetting and slicing dataframs can be weird.
when you subset a `data.frame` that ends up with 1 column it will come back as a `vector` and not a `data.frame`

Note, `tibble` objects don't do this.

You can prevent this automatic conversion to a vector by passing in the `drop = FALSE` into the `[ ]` subsetter

```{r}
mtcars[mtcars$mpg < 15.43, 'cyl', drop = FALSE]
```

Otherwise you will have to manually convert the vector back to a `data.frame` by using `as.data.frame`

```{r}
as.data.frame(mtcars[mtcars$mpg < 15.43, 'cyl'])
```

But then you'll have to end up changing the column name

```{r}
bool_df <- as.data.frame(mtcars[mtcars$mpg < 15.43, 'cyl'])
names(bool_df) <- 'cyl'
bool_df
```

## Multiple boolean conditions

You can use the `|` or `&` to have an OR or AND operator in your boolean condition, respectively.

```{r}
mtcars[mtcars$mpg < 15.43 | mtcars$cyl == 4, ]
```

```{r}
mtcars[mtcars$mpg < 15.43 & mtcars$cyl == 4, ]
```


# NA Missing Values

Missing values in R are coded as an `NA` value

```{r}
NA
```


You can't use the `==` operator to check for missing values

```{r}
NA == NA
```

```{r}
NA == TRUE
```

```{r}
NA == FALSE
```

```{r}
NA == 3
```


You have to use a special function to check for missing values

```{r}
is.na(NA)
```