Data type optimization in Data Analysis Python - Time & Space Complexity

Time Complexity: Data type optimization

O(n)

Understanding Time Complexity

When we change data types in data analysis, it can affect how fast our code runs.

We want to know how these changes affect the time it takes to process data.

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import pandas as pd

data = pd.DataFrame({
    'numbers': range(1000000),
    'floats': [float(x) for x in range(1000000)]
})

data['numbers'] = data['numbers'].astype('int32')  # change to smaller integer type
result = data['numbers'].sum()

This code changes a column's data type to a smaller integer type and then sums the values.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Summing all values in the column.
• How many times: Once for each of the 1,000,000 rows.

How Execution Grows With Input

As the number of rows grows, the sum operation takes longer because it looks at each value once.

Input Size (n)	Approx. Operations
10	10 sums
100	100 sums
1000	1000 sums

Pattern observation: The time grows directly with the number of rows.

Final Time Complexity

Time Complexity: O(n)

This means the time to sum grows in a straight line as the data size grows.

Common Mistake

[X] Wrong: "Changing data types always makes the code run faster."

[OK] Correct: Changing data types can save memory but does not change how many times the code must add values.

Interview Connect

Understanding how data size and types affect speed shows you can write efficient data analysis code.

Self-Check

"What if we changed the sum operation to a nested loop over the data? How would the time complexity change?"

Practice

1. Why is data type optimization important when working with large datasets in pandas?

easy

A. It sorts the data in ascending order.

B. It changes the data values to strings.

C. It deletes unnecessary columns automatically.

D. It reduces memory usage and speeds up data processing.

Solution

1. Step 1: Understand data type optimization purpose

   Data type optimization aims to reduce the memory used by data and improve processing speed.

2. Step 2: Identify the correct benefit

   Reducing memory and speeding up processing matches the purpose, unlike changing values or sorting.

3. Final Answer:

   It reduces memory usage and speeds up data processing. -> Option D

4. Quick Check:

   Data type optimization = reduce memory and speed up [OK]

Hint: Think about saving memory and faster work [OK]

Common Mistakes:

• Confusing optimization with sorting or deleting data
• Thinking it changes data values instead of types
• Assuming it automatically cleans data

2. Which pandas method is used to change the data type of a DataFrame column?

easy

A. astype()

B. convert()

C. change_type()

D. modify_dtype()

Solution

1. Step 1: Recall pandas method for type conversion

   The pandas method to change a column's data type is astype().

2. Step 2: Verify other options

   convert(), change_type(), and modify_dtype() are not valid pandas methods.

3. Final Answer:

   astype() -> Option A

4. Quick Check:

   pandas type change = astype() [OK]

Hint: Remember pandas astype() for type changes [OK]

Common Mistakes:

• Using non-existent methods like convert()
• Confusing with methods for data cleaning
• Trying to assign types directly without astype()

3. Given the code below, what will be the memory usage of the DataFrame after optimization?

import pandas as pd
import numpy as np

df = pd.DataFrame({
  'A': np.arange(1000, dtype='int64'),
  'B': ['cat', 'dog', 'mouse', 'cat'] * 250
})

memory_before = df.memory_usage(deep=True).sum()
df['B'] = df['B'].astype('category')
memory_after = df.memory_usage(deep=True).sum()

print(memory_after < memory_before)

medium

A. True, memory usage decreases after converting to category.

B. False, memory usage increases after converting to category.

C. True, memory usage stays the same after converting to category.

D. Error, astype('category') is invalid.

Solution

1. Step 1: Understand memory usage before optimization

   Column 'B' is object type with repeated strings, using more memory.

2. Step 2: Effect of converting to category

   Converting 'B' to category stores unique values once, reducing memory.

3. Final Answer:

   True, memory usage decreases after converting to category. -> Option A

4. Quick Check:

   Category type reduces memory = True [OK]

Hint: Categorical reduces repeated string memory [OK]

Common Mistakes:

• Thinking category increases memory
• Assuming memory stays same after astype
• Believing astype('category') causes error

4. Identify the error in the following code snippet that tries to optimize data types:

import pandas as pd

df = pd.DataFrame({'col': [1, 2, 3, 4]})
df['col'] = df['col'].astype('int8')
df['col'] = df['col'].astype('category')
df['col'] = df['col'].astype('int16')

medium

A. int8 is not a valid data type in pandas.

B. Cannot convert from category back to int16 directly.

C. astype() cannot be used multiple times on the same column.

D. DataFrame creation syntax is incorrect.

Solution

1. Step 1: Check data type conversions

   Converting from int8 to category is valid, but category stores codes, not original ints.

2. Step 2: Conversion from category to int16

   Directly converting category to int16 causes error because category is not numeric type.

3. Final Answer:

   Cannot convert from category back to int16 directly. -> Option B

4. Quick Check:

   Category to int conversion invalid [OK]

Hint: Category to int needs .cat.codes first [OK]

Common Mistakes:

• Assuming int8 is invalid
• Thinking astype can chain without issues
• Believing DataFrame creation is wrong

5. You have a DataFrame with a column 'status' containing only three unique string values repeated many times. Which is the best way to optimize memory usage for this column while keeping it easy to analyze?

hard

A. Leave 'status' as object type because conversion wastes time.

B. Convert 'status' to integer type directly with astype('int32').

C. Convert 'status' to a categorical type using astype('category').

D. Convert 'status' to float type with astype('float64').

Solution

1. Step 1: Analyze the column data

   'status' has few unique strings repeated many times, ideal for categorical type.

2. Step 2: Choose the best optimization

   astype('category') reduces memory by storing unique values once and codes for repeats.

3. Final Answer:

   Convert 'status' to a categorical type using astype('category'). -> Option C

4. Quick Check:

   Few unique strings = use category [OK]

Hint: Few unique strings? Use category type [OK]

Common Mistakes:

• Trying to convert strings directly to int or float
• Ignoring memory benefits of categorical
• Thinking object type is always best