Hints

• Use the following slicing trick to copy a list into a new variable.

>>> copyOfListOfCars = listOfCars[:]

• Create new variables to hold intermediate lists.

>>> list1 = ['Rover', 'Infiniti', 'Honda', 'Kia']
>>> list2 = list1[::2]
>>> list3 = random.sample(list2, 1)

After you have spent time solving the challenge (or trying but failing miserably -- that's okay), then review the solution below.

Solution

To start tackling the challenge, I used the numbered steps as a coding guide, commenting-out the steps with the # symbol (called "hash", "pound", or "octothorpe").

Phase 1

# 1. loads the 'random' module
import random

# 2. creates a list of ten cookie flavors
recipes = ['molasses', 'ginger', 'chocolate', 'samoa',
    'no bake', 'mint', 'sugar', 'peanut', 'oatmeal', 'bacon']

print recipes

# 3. makes a copy of the list.

# 4. randomizes the list.

# 5. slices the randomized list in half.

# 6. gets 3 interval samples from the sliced+random list.

# 7. prints the final list, you sample.

Next, below each commented line, I typed in the code to satisfy that specific step, testing the script after every command I wrote.

Phase 2

# 1. loads the 'random' module
import random

# 2. creates a list of ten cookie flavors
recipes = ['molasses', 'ginger', 'chocolate', 'samoa',
    'no bake', 'mint', 'sugar', 'peanut', 'oatmeal', 'bacon']

# 3. makes a copy of the list.
recipeCopy = recipes[:]

print recipeCopy

# 4. randomizes the list.

# 5. slices the randomized list in half.

# 6. gets 3 interval samples from the sliced+random list.

# 7. prints the final list, you sample.

Phase 3

# 1. loads the 'random' module
import random

# 2. creates a list of ten cookie flavors
recipes = ['molasses', 'ginger', 'chocolate', 'samoa',
    'no bake', 'mint', 'sugar', 'peanut', 'oatmeal', 'bacon']

# 3. makes a copy of the list.
recipeCopy = recipes[:]

# 4. randomizes the list.
recipeRandom = random.sample(recipeCopy, len(recipeCopy))

print recipeRandom

# 5. slices the randomized list in half.

# 6. gets 3 interval samples from the sliced+random list.

# 7. prints the final list, you sample.

Phase 4, the end

# 1. loads the 'random' module
import random

# 2. creates a list of ten cookie flavors
recipes = ['molasses', 'ginger', 'chocolate', 'samoa',
    'no bake', 'mint', 'sugar', 'peanut', 'oatmeal', 'bacon']

# 3. makes a copy of the list.
recipeCopy = recipes[:]

# 4. randomizes the list.
recipeRandom = random.sample(recipeCopy, 10)

# 5. slices the randomized list in half.
recipeHalf = recipeRandom[:6]

# 6. gets 3 interval samples from the sliced+random list.
sample = recipeHalf[::2]

# 7. prints the final list, you sample.
print sample

Output:

['ginger', 'mint', 'peanut']

Sampling Summary

The approaches to sampling that you've seen so far get more useful as the amount of data increases. If you are given 100,000 records of political campaign donations, it may be unmanageable to review each one. But by sampling every 10th donation, for example, the number of donations through which to shift is reduced dramatically while still preserving much of the 'character' if the data. That is, the sample may be suitable enough for basic analysis. If you find something interesting with the sample, such as one political party receiving much more funding than another, you can then take more samples of the original data set to see if your finding holds true across all the samples, and the population as a whole. Statisticians call the concept of studying a sample to see if it applies to the whole inferential statistics. You can infer characteristics about a population by studying samples that are representative. This is both an art and a science, and is a field of statistics unto itself.

You can perhaps infer that your recent batch of baked cookies are tasty by randomly sampling just a couple cookies, perhaps one from the edge of the pan and one from the center of the pan.

Just like it would be foolish to eat all of the cookies that you baked for other people, it's often infeasible to collect data on every person or thing that you want to study.

Suppose you are interested in studying people diagnosed with schizophrenia. It would be a monumental task to try locating all schizophrenic people. Instead, what you or a medical researcher may do is to study schizophrenia people who are served by a particular hospital, and then carefully draw conclusions about the population as a whole. These conclusions may then be supported by a different researcher who studied schizophrenia at a different hospital.