pandas

Question 1

Given the following dictionary, figure out the following?

sports = { ‘Archery’ : ‘Bhutan’,

‘Golf’ : ‘Scotland’,

‘Sumo’ ; ‘Japan’,

‘Taekwondo’ : ‘South Korea’ }

• How do you get the Series?
• How do you get the 4th value?
• How do you get the value of ‘Golf’?

Answer 1

import pandas as pd

s = pd.Series(sports)

print(s)

1. Output attached
2. s.iloc[3] or s[3]
3. s.loc[‘Golf’] or s[‘Golf’]

Question 2

• The following purchases took place:
  
  • Purchase 1: Chris purchased dog food for 22.50
  • Purchase 2: Kevyn purchased kittly litter for 2.50
  • Purchase 3: Vinod purchased bird seed for 5.00

1. Place these purchases in a Series.
2. Place these Series within a DataFrame with indexes
   
   1. Give Purchase 1 and 2 to Index “Store 1”
   2. Give Purchase 3 to Index “Store 2”

Answer 2

• See attached

Question 3

Given the dataframe ‘df’ composed of three purchases, how would you:

1. Get the values stored within the index ‘Store 2’?
2. Get the list of all items that have been purchased, regardless of where they were purchased?
3. Get the COST of items purchased at ‘Store 1’?
4. Return the NAME and COST for all items from all stores?
5. Show the dataframe without the ‘Store 1’ indexed data?
6. Delete the ‘Name’ values from the dataframe?
7. Insert a new column into the dataframe named ‘Location’?
8. Update the dataframe by applying a discount of 20% across all values in the ‘Cost’ column?

Answer 3

1. df.loc[‘Store 2’]
2. df[‘Item Purchased’]
3. df.loc[‘Store 2’, ‘Cost’]
4. df.loc[: , [‘Name’, ‘Cost’] ]
5. df.drop(‘Store 1’)
   
   • NOTE: This creates a copy with the Store 1 index removed, and does not actually remove Store 1 index values from the original dataframe
6. del df[‘Name’]
   
   • NOTE: This permanently deletes this data from the dataframe
7. df[‘Location’] = None
8. df[‘Cost’] *= .80; print(df)

Question 4

What practice is the following an example of?

df.loc[‘Store’][‘Cost’]

• Why should you not do it?
• What would be a different way to do this?

Answer 4

• This is referred to as chaining
  
  • this should not be done because it creates a copy of the dataframe instead of a view and can cause unpredicable results
  • df.loc[‘Store’, ‘Cost’]

Question 5

How do you create a new series named ‘costs’ that is based on the original series ‘Cost’ data?

• Using broadcasting, how do you increase the cost of each item by 2?
• What will happen to the ‘Cost’ values in the original dataframe?
  
  • What should you do if this is not your intention?

Answer 5

• costs = df[‘Cost’]
• costs += 2
  
  • The costs in the original dataframe will also increase by 2
  • If this is not the desired outcome (e.g. if you do not want to change the original dataframe values) then use the copy method

Question 6

Given a csv file named ‘olympics.csv’, how would you:

• Read the file into Pandas
  
  • Make the first column the index column
  • Skp the first row
• Print the first 5 rows
• Rename the column headers as follows:
  
  • If the column is named ‘01’, rename to ‘Gold’
  • If the column is named ‘02’, rename to ‘Silver’
  • If the column is named ‘03’, rename to ‘Bronze’

Answer 6

• df = pd.read_csv(‘olympics.csv’, index_col=0, skiprows=1)
• df.head()
• for col in df.columns:

if col[:2] == ‘01’:

df.rename(columns={col:’Gold’]}, inplace = True)

if col[:2] == ‘02’:

df.rename(columns={col:’Silver’]}, inplace = True)

if col[:2] == ‘03’:

df.rename(columns={col:’Bronze’]}, inplace = True)

Question 7

• Using boolean masking, how would you return values from a dataframe where values in the ‘Gold’ column are greater than 0?
• Using boolean masking, create a new series composed of data from the original dataframe where the values in the ‘Gold’ column are greater than 0?
  
  • Return how many records are in this new dataframe?

Answer 7

• df[‘Gold’] > 0
• only_gold = df.where(df[‘Gold’] > 0)
• df[‘Gold’].count()

Question 8

Given the dataframe of purchases, return the NAMES of customers whose purchases COST more than $3.00

Answer 8

• df[‘Name’][df[‘Cost’]>3]

Question 9

What Pandas function takes a lsit of columns and promotes those columns to an index?

• When this function is used, what happens to the original index (by default)?
• Example from olympic.csv data:
  
  1. Create a new index of ‘Gold’
  2. Make sure to preserve the original index ‘Country’
  3. Reset the index to remove the unnecessary header

Answer 9

• set_index
  
  • This is a destructive process, meaning it does not keep the original index
• Example:
  
  1. df[‘country’] = df.index
  2. df = df.set_index(‘Gold’)
  3. df = df.reset_index()

Question 10

Given the following DataFrame, carry out the following tasks:

• df = pd.DataFrame([purchase_1, purchase_2, purchase_3], index=[‘Store 1’, ‘Store 1’, ‘Store 2’]

1. Reindex the purchase records DataFrame to be indexed hierarchically, first by store, then by person.
2. Name these indexes ‘Location’ and ‘Name’.
3. Then add a new entry to it with the value of:

• Name: ‘Kevyn’
• Item Purchased: ‘Kitty Food’
• Cost: 3.00
• Location: ‘Store 2’

Answer 10

1. Set the new index of ‘Name’ (the store index already exists)
   
   1. df = df.set_index([df.index, ‘Name’])
2. Name the indexes
   
   1. df.index.names = [‘Location’, ‘Name’]
3. df = df.append(pd.Series(data={‘Cost’: 3.00, ‘Item Purchased’: ‘Kitty Food’}, name=(‘Store 2’, ‘Kevyn’)))