### Dataset:

cars.okc (See the raw data)

The Cars dataset has 392 observations and 7 variables describing attributes of cars.
The complete variables are:

1. Miles Per Gallon
2. Cylinders (3, 4, 5, 6, and 8)
3. Horsepower
4. Weight
5. Acceleration
6. Year of manufacture (70-90)
7. Country of Origin (1=American, 2=European, 3=Japanese)

### Hypotheses:

1. The cars with high MPG will be mostly 4 cylinder cars with low weight and acceleration.
2. American makes heavier cars than does Japanese.

### Analysis (to verify or refute any of these hypothese):

1. First, deselect Year and Origin (i.e. only selecting MPG, Cylinders, Horsepower, Weight, Acceleration).

Then, highlight cars with high MPG and see that most of them are 4 cylinder cars with low weight, but, not necessary low acceleration. Therefore, refute this hypothesis.

Note that a couple of exceptions exist, and by moving the brush to exclude 4 cylinder cars, we see data points corresponding to 5 and 6 cylinder cars that have good fuel economy.

2. First, select only Weight and Origin (and deselecting the other variables).

Then, highlight American (Origin=1) and Japanese (Origin=3) on parallel coordinates. The graphs show that Japanese makes light weight cars, and American makes all range (from light to heavy) weight cars.

When brushing heavier cars (Weight>3000) on scatter plot matrix, it shows that all data points higher than 3000 on Weight are not mapped to Japanese. That is, heavy cars are mostly from American.

When brushing American (Origin=1), American also makes light weight cars. Therefore, refute this hypothesis.

Training