Category codes and labels in Pandas - Step-by-Step Execution

Concept Flow - Category codes and labels

Create categorical data

↓

Assign categories with labels

↓

Access category codes

↓

Use codes for analysis or storage

↓

Map codes back to labels if needed

This flow shows how to create categorical data, assign labels, access their integer codes, and use them for efficient analysis.

Execution Sample

Pandas

import pandas as pd
colors = pd.Series(['red', 'blue', 'green', 'blue', 'red'])
cats = colors.astype('category')
codes = cats.cat.codes
labels = cats.cat.categories

This code creates a categorical series from color names, then extracts integer codes and category labels.

Execution Table

Step	Action	Data/Variable	Result/Value
1	Create Series with colors	colors	0:red,1:blue,2:green,3:blue,4:red
2	Convert to categorical	cats	Categories: ['blue', 'green', 'red'] with codes [2,0,1,0,2]
3	Access category codes	codes	[2,0,1,0,2]
4	Access category labels	labels	Index(['blue', 'green', 'red'])
5	Use codes for indexing or storage	codes	Integer array representing categories
6	Map codes back to labels	codes + labels	2->red,0->blue,1->green
7	End		All categories and codes assigned correctly

💡 All categorical codes and labels extracted and mapped successfully

Variable Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	Final
colors	Empty	['red','blue','green','blue','red']	Same	Same	Same
cats	N/A	Categorical with categories ['blue','green','red']	Same	Same	Same
codes	N/A	N/A	[2,0,1,0,2]	Same	Same
labels	N/A	N/A	N/A	['blue','green','red']	Same

Key Moments - 2 Insights

Why do the category codes not match the original order of colors?

Category codes are assigned based on sorted unique categories (alphabetical order), not the original order. See execution_table step 2 where categories are ['blue', 'green', 'red'] sorted alphabetically.

How can I get the original color names back from the codes?

Use the category labels with the codes to map back. For example, code 0 corresponds to 'blue' as shown in execution_table step 6.

Visual Quiz - 3 Questions

Look at the execution_table at step 3, what is the code for the color 'green'?

A 1

B 2

C 0

D 3

💡 Hint

Check the codes array in step 3 and the category order in step 2.

At which step do we convert the original Series into categorical data?

A Step 3

B Step 1

C Step 2

D Step 4

💡 Hint

Look for the action 'Convert to categorical' in the execution_table.

If a new color 'yellow' is added, what happens to the category codes?

A All codes shift by one

B Codes remain the same for existing colors, 'yellow' gets a new code

C Codes become strings instead of integers

D Codes reset to zero for all colors

💡 Hint

Think about how pandas assigns codes only to new categories without changing existing ones.

Concept Snapshot

Category codes and labels in pandas:
- Convert a Series to 'category' dtype.
- Categories are sorted unique values.
- Codes are integer indexes for categories.
- Use .cat.codes for codes, .cat.categories for labels.
- Codes help efficient storage and analysis.

Full Transcript

We start by creating a pandas Series with color names. Then, we convert this Series to a categorical type, which organizes unique colors as categories sorted alphabetically. Each category gets an integer code starting from zero. We extract these codes using .cat.codes and the category labels using .cat.categories. These codes represent the original data efficiently and can be mapped back to the labels when needed. This process helps in saving memory and speeding up analysis when working with repeated categorical data.

Practice

1. What does cat.codes show in a pandas categorical Series?

easy

A. The data type of the Series

B. The numeric codes representing each category

C. The count of each category

D. The text labels of the categories

Solution

1. Step 1: Understand what cat.codes returns

   cat.codes gives the integer codes for each category in the Series, replacing repeated text with numbers.

2. Step 2: Differentiate from cat.categories

   cat.categories shows the text labels, not the codes.

3. Final Answer:

   The numeric codes representing each category -> Option B

4. Quick Check:

   cat.codes = numeric codes [OK]

Hint: Remember: codes are numbers, categories are labels [OK]

Common Mistakes:

• Confusing codes with category labels
• Thinking codes show counts
• Assuming codes show data type

2. Which of the following is the correct way to access the category labels of a pandas categorical Series s?

easy

A. s.cat.categories

B. s.cat.codes

C. s.categories

D. s.cat.labels

Solution

1. Step 1: Recall the attribute for category labels

   The category labels are accessed via cat.categories on the Series.

2. Step 2: Check the options for correct syntax

   s.cat.categories is the correct syntax; s.categories is invalid, and cat.labels does not exist.

3. Final Answer:

   s.cat.categories -> Option A

4. Quick Check:

   Category labels = s.cat.categories [OK]

Hint: Use cat.categories to get labels, not just categories [OK]

Common Mistakes:

• Using s.categories instead of s.cat.categories
• Confusing cat.codes with labels
• Trying to use non-existent cat.labels

3. Given the code:

import pandas as pd
s = pd.Series(['apple', 'banana', 'apple', 'orange'], dtype='category')
print(s.cat.codes.tolist())

What is the output?

medium

A. [0, 1, 0, 2]

B. [1, 2, 1, 3]

C. [0, 0, 1, 2]

D. [1, 0, 1, 2]

Solution

1. Step 1: Identify categories and their order

   Categories are sorted alphabetically by default: ['apple', 'banana', 'orange'] with codes 0,1,2 respectively.

2. Step 2: Map each value to its code

   'apple' -> 0, 'banana' -> 1, 'apple' -> 0, 'orange' -> 2, so the list is [0, 1, 0, 2].

3. Final Answer:

   [0, 1, 0, 2] -> Option A

4. Quick Check:

   Alphabetical categories -> codes match [OK]

Hint: Categories sorted alphabetically by default [OK]

Common Mistakes:

• Assuming codes start at 1 instead of 0
• Ignoring alphabetical sorting of categories
• Mixing up category order

4. Which of the following is correct about this code snippet?

import pandas as pd
s = pd.Series(['cat', 'dog', 'cat'], dtype='category')
s.cat.categories = ['dog', 'cat']
print(s.cat.codes.tolist())

medium

A. You cannot assign new categories with different order directly

B. The categories list must be sorted alphabetically

C. You must convert Series to string before assigning categories

D. The code will run without errors and print [0, 1, 0]

Solution

1. Step 1: Understand category assignment rules

   Assigning to cat.categories redefines the labels for existing codes without changing the codes or raising errors, as long as new list has same length and unique values.

2. Step 2: Trace execution

   Original categories ['cat', 'dog'], codes [0,1,0]. New ['dog', 'cat'] same length/unique, codes unchanged, prints [0, 1, 0].

3. Final Answer:

   The code will run without errors and print [0, 1, 0] -> Option D

4. Quick Check:

   New categories redefine code labels, codes unchanged [OK]

Hint: Reassigning cat.categories changes labels but keeps codes intact [OK]

Common Mistakes:

• Thinking reordering causes ValueError
• Expecting codes to adjust to new order like [1, 0, 1]
• Assuming categories must stay alphabetical

5. You have a pandas categorical Series s with categories ['low', 'medium', 'high']. You want to rename the categories to ['L', 'M', 'H'] without changing the data. Which code correctly returns the renamed Series?

hard

A. s.cat.categories = ['L', 'M', 'H']

B. s.cat.categories.rename(['L', 'M', 'H'])

C. s.cat.rename_categories(['L', 'M', 'H'])

D. s.rename_categories(['L', 'M', 'H'])

Solution

1. Step 1: Understand how to rename categories

   The method cat.rename_categories() returns a new Series with renamed categories without changing data codes.

2. Step 2: Check other options for correctness

   Direct assignment to cat.categories works but can cause errors if not careful; cat.categories.rename() is invalid; rename_categories() is not a Series method.

3. Final Answer:

   s.cat.rename_categories(['L', 'M', 'H']) -> Option C

4. Quick Check:

   Use cat.rename_categories() to rename safely [OK]

Hint: Use cat.rename_categories() to rename categories safely [OK]

Common Mistakes:

• Trying to rename categories by invalid methods
• Directly assigning categories without matching length
• Confusing Series methods with cat accessor methods