Learn Making Predictions with scikit-learn | Data-Driven Decision Making

Predictive modeling gives you the power to use past data to forecast future outcomes, a capability that can be transformative for startups. Whether you want to estimate next month's sales, predict which customers might leave (churn), or determine the impact of your marketing spend, predictive models help you make smarter, data-driven decisions. By leveraging these models, you can allocate resources more effectively, anticipate challenges, and spot new opportunities before your competitors.


              12345678910111213141516
            
import numpy as np
from sklearn.linear_model import LinearRegression

# Example data: advertising spend vs. sales
# X: advertising spend in thousands of dollars
# y: sales in thousands of units
X = np.array([[5], [10], [15], [20], [25]])  # advertising spend
y = np.array([7, 9, 15, 18, 21])            # sales

# Create and fit the linear regression model
model = LinearRegression()
model.fit(X, y)

# Print the learned coefficients
print("Intercept:", model.intercept_)
print("Slope:", model.coef_[0])

When you fit a linear regression model using scikit-learn, you are essentially finding the best straight line that describes the relationship between your input variable (for example, advertising spend) and your target variable (such as sales). The model learns two key values: the intercept (where the line crosses the y-axis) and the slope (how much the target variable changes for each unit increase in the input). Interpreting these results helps you understand how strongly your input influences your target, and gives you a formula to make future predictions.


              123456789
            
# Predict sales for new advertising spend values
new_spend = np.array([[12], [18], [30]])  # new advertising spend
predicted_sales = model.predict(new_spend)

# Compare predictions with actual values (if available)
actual_sales = np.array([11, 17, 26])  # hypothetical actual sales

for spend, pred, actual in zip(new_spend.flatten(), predicted_sales, actual_sales):
    print(f"Ad Spend: ${spend}k - Predicted Sales: {pred:.1f}k, Actual Sales: {actual}k")

Everything was clear?

Thanks for your feedback!

Section 2. Chapter 5

Ask AI

Ask anything or try one of the suggested questions to begin our chat

Suggested prompts:

Can you explain how to interpret the predicted sales versus actual sales results?

What are some ways to improve the accuracy of this predictive model?

How can I use this model to make predictions for other types of data?

Swipe to show menu


              12345678910111213141516
            
import numpy as np
from sklearn.linear_model import LinearRegression

# Example data: advertising spend vs. sales
# X: advertising spend in thousands of dollars
# y: sales in thousands of units
X = np.array([[5], [10], [15], [20], [25]])  # advertising spend
y = np.array([7, 9, 15, 18, 21])            # sales

# Create and fit the linear regression model
model = LinearRegression()
model.fit(X, y)

# Print the learned coefficients
print("Intercept:", model.intercept_)
print("Slope:", model.coef_[0])


              123456789
            
# Predict sales for new advertising spend values
new_spend = np.array([[12], [18], [30]])  # new advertising spend
predicted_sales = model.predict(new_spend)

# Compare predictions with actual values (if available)
actual_sales = np.array([11, 17, 26])  # hypothetical actual sales

for spend, pred, actual in zip(new_spend.flatten(), predicted_sales, actual_sales):
    print(f"Ad Spend: ${spend}k - Predicted Sales: {pred:.1f}k, Actual Sales: {actual}k")

Everything was clear?

Thanks for your feedback!

Section 2. Chapter 5

Making Predictions with scikit-learn

1. What is the purpose of a predictive model in business?

2. Which scikit-learn class is used for linear regression?

3. Why is it important to compare predictions to actual results?

Making Predictions with scikit-learn

1. What is the purpose of a predictive model in business?

2. Which scikit-learn class is used for linear regression?

3. Why is it important to compare predictions to actual results?