Data Analysis-Intro to wdcplot-mtcars Example.Rmd

# Intro to wdcplot; mtcars Example
===
***
When in edit mode, run the analysis using the "play" (arrow) button in the navigation bar.

## Let's load the popular mtcars dataset. It comes pre-bundled with R.
```{r}
data(mtcars)
```

## What does the data look like?
```{r}
head(mtcars)
```
# Looks like mtcars is a data frame:
```{r}
class(mtcars)
```
## So the structure is ...

```{r}
str(mtcars)
```
## What are the total number of rows?
```{r}
nrow(mtcars)
```
## How do we access a column?
```{r}
head(mtcars$mpg)
```
## What's the type/class?
```{r}
class(mtcars$mpg)

## OR
head(mtcars[,1])

## OR
head(mtcars[, "mpg"])
```
## We don't want a numeric vector. We want a data.frame slice.

```{r}
head(mtcars["mpg"])
```
## Did we do it right? What's the type/class?

```{r}
class(mtcars["mpg"])
```
## How do I get more than one column?

```{r}
head(mtcars[c("mpg","hp")])
```
## How do we know the row names?

```{r}
row.names(mtcars)
```
## How do we access a specific value?

```{r}
mtcars["Mazda RX4", "cyl"] 

# OR
mtcars[1,2]
```
## How do I access a row?

```{r}
mtcars["Mazda RX4", ] 
```
## Also the same as ...
```{r}
mtcars[1, ]
```
## Vehicles with automatic transmission?

```{r}
head(mtcars[mtcars$am==0,])
```
## Mileage for automatic transmission vehicles?

```{r}
mtcars[mtcars$am==0,]$mpg
```
## Let's know our data better.

```{r}
summary(mtcars)
```
## We don't want to keep typing mtcars$... so we use attach:

```{r}
attach(mtcars)
```

## You can now directly refer mtcars$mpg as mpg.

```{r}
identical(mtcars$mpg, mpg)
```
## Let's summarize our data:

```{r}
table(cyl)
```
## Plotting is easy...

## bar plot
```{r}
barplot(mpg)
```
## OR a histogram
```{r}
hist(mpg)

```
## OR a boxplot
```{r}
boxplot(mpg)
```
## Looks like mpg are pretty low:

```{r}
mean(mpg)
```
## Mean of mileage for 4 cylinder cars:

```{r}
mean(mpg[cyl==4])
```
## vs 8 cylinder cars:

```{r}
mean(mpg[cyl==8])
```
## Is that because of a higher number of cylinders?

```{r}

plot(cyl, mpg)

```
## Let's fit a regression line:

```{r}
lm(mpg~cyl+hp)


```
## Pearson correlation, is that a good fit?

```{r}
cor(mpg,cyl)^2
cor(mpg, hp)^2
```
## Let's plot based on number of cylinders

```{r}
plot(hp,mpg,pch=19, col=cyl)
# add a legend
legend(250, 30, pch=19, col=c(4,6,8), legend=c("4 cylinders","6 cylinders","8 cylinders"))  
```
## Now we do the cool stuff:
```{r fig.width=11, fig.height=11, tidy=FALSE}
require(rcloud.dcplot)
wdcplot(mtcars, 
dimensions(..index.., mpg, wt, cyl),
groups(mpgGroup = group(mpg, bin(2)),
       wtGroup = group(wt, group = bin(0.5)),
       cylinders = group(cyl)),
charts(SWvL = bubble('Miles Per Gallon vs. Weight',
                     dimension = ..index..,
                     color = cyl,  
                     x = wt,
                     y = mpg,
                     r = ..selected.. * 3,
                     label = NULL),
       cyl = pie("Cylinders",group = cylinders),
       PW = bar('Miles Per Gallon', group = mpgGroup, x.domain = c(10.4,33.9)),
       PL = bar('Weight', group = wtGroup, x.domain = c(1.5,5.5))
))
```
```{r}
detach(mtcars)
```