NumPy

Numerical Computations

Use cases: Perform numerical calculations on large datasets like sales trends. This is used for efficiently handling numerical data for forecasting and product insights.

Breakdown of the below code is as follows:

• np .array([...]) in which np calls the library and .array([...]) converts a python list into a NumPy array like [1000, 1200, 900, 1500] is a list of revenue values and the revenue stores the array in the variable.

• A NumPy array is a special type if list that: Stores numbers more efficiently than a regular Python list and allows for fast mathematical operations.

• .mean() calculates the average of all numbers in the revenue array

import numpy as np

revenue = np.array([1000, 1200, 900, 1500])
avg_revenue = np.mean(revenue)  # Calculate average revenue
print(avg_revenue)

Table: NumPy Functions

NumPy Function	Purpose in A/B Testing	Example Use Case
np.mean()	Calculate the average of a numerical column	Find average clicks per user
np.median()	Get the middle value of sorted data	Find median clicks per user
np.std()	Calculate standard deviation	Check variation in user clicks
np.var()	Compute variance	Measure spread of user engagement
np.percentile()	Get percentile values (25th, 50th, 75th)	Understand user behavior distribution
np.min() / np.max()	Get minimum and maximum values	Identify lowest and highest clicks
np.random.choice()	Randomly select a value from an array	Simulate user interactions

Syntax references for NumPy functions are as below:

np.mean(array)  # Compute mean
np.median(array)  # Compute median
np.std(array)  # Compute standard deviation
np.var(array)  # Compute variance
np.min(array)  # Get minimum value
np.max(array)  # Get maximum value
np.percentile(array, 25)  # Get 25th percentile value