AI shopping just beat search at its own game on Prime Day

Learn to build a simple AI-powered shopping recommendation system that mimics Amazon's Prime Day technology using Python and basic machine learning concepts.

Introduction

In this tutorial, you'll learn how to create a simple AI-powered shopping recommendation system using Python and basic machine learning concepts. This tutorial mirrors the technology that helped Amazon's Prime Day achieve record-breaking sales by using AI to recommend products to customers. We'll build a system that can suggest products based on user preferences and shopping history.

Prerequisites

To follow this tutorial, you'll need:

A computer with Python installed (version 3.6 or higher)
Basic understanding of Python programming concepts
Internet connection for downloading required packages

Step-by-Step Instructions

Step 1: Set Up Your Python Environment

First, we need to create a new Python project and install the necessary libraries. Open your terminal or command prompt and run the following commands:

mkdir ai-shopping-recommender
 cd ai-shopping-recommender
 python -m venv shopping_env
 source shopping_env/bin/activate  # On Windows use: shopping_env\Scripts\activate
 pip install pandas scikit-learn

Why we do this: We're creating a separate environment to avoid conflicts with other Python projects. The libraries we install will help us process data and build our recommendation model.

Step 2: Create Sample Product Data

Next, we'll create a simple dataset of products. Create a file called products.py:

import pandas as pd

def create_sample_products():
    products = {
        'product_id': [1, 2, 3, 4, 5, 6, 7, 8],
        'name': ['Wireless Headphones', 'Smartphone', 'Laptop', 'Smart Watch', 'Tablet', 'Camera', 'Speaker', 'Gaming Console'],
        'category': ['Electronics', 'Electronics', 'Electronics', 'Electronics', 'Electronics', 'Electronics', 'Electronics', 'Electronics'],
        'price': [89.99, 699.99, 1299.99, 249.99, 399.99, 599.99, 129.99, 499.99],
        'rating': [4.5, 4.7, 4.8, 4.3, 4.6, 4.4, 4.2, 4.9]
    }
    return pd.DataFrame(products)

if __name__ == '__main__':
    df = create_sample_products()
    print(df)

Why we do this: This creates a realistic dataset of products that we can use to train our recommendation system. The data includes key attributes that users might consider when shopping.

Step 3: Create User Preference Data

Now, let's create a user profile system. Create a file called users.py:

import pandas as pd

def create_sample_users():
    users = {
        'user_id': [1, 2, 3],
        'preferred_category': ['Electronics', 'Electronics', 'Electronics'],
        'max_budget': [1000, 2000, 500],
        'min_rating': [4.0, 4.5, 4.0]
    }
    return pd.DataFrame(users)

if __name__ == '__main__':
    df = create_sample_users()
    print(df)

Why we do this: Users have different preferences, budgets, and rating requirements. This data helps our AI understand what each user is likely to buy.

Step 4: Build the Recommendation Engine

Now, let's create the core recommendation system. Create a file called recommender.py:

import pandas as pd
from sklearn.metrics.pairwise import cosine_similarity


class ShoppingRecommender:
    def __init__(self, products_df, users_df):
        self.products = products_df
        self.users = users_df

    def recommend_products(self, user_id, num_recommendations=3):
        # Get user preferences
        user = self.users[self.users['user_id'] == user_id].iloc[0]
        
        # Filter products based on user preferences
        filtered_products = self.products[
            (self.products['price'] <= user['max_budget']) &
            (self.products['rating'] >= user['min_rating'])
        ]
        
        # Sort by rating (highest first)
        recommended = filtered_products.sort_values('rating', ascending=False)
        
        # Return top recommendations
        return recommended.head(num_recommendations)

    def get_similar_products(self, product_id, num_similar=3):
        # Simple similarity based on price and rating
        target_product = self.products[self.products['product_id'] == product_id].iloc[0]
        
        # Calculate similarity scores
        self.products['similarity'] = (
            (self.products['price'] - target_product['price']).abs() * -1 +
            (self.products['rating'] - target_product['rating']) * 10
        )
        
        # Get similar products
        similar = self.products[self.products['product_id'] != product_id].sort_values('similarity', ascending=False)
        return similar.head(num_similar)

if __name__ == '__main__':
    # Load data
    from products import create_sample_products
    from users import create_sample_users
    
    products_df = create_sample_products()
    users_df = create_sample_users()
    
    # Create recommender
    recommender = ShoppingRecommender(products_df, users_df)
    
    # Get recommendations for user 1
    recommendations = recommender.recommend_products(1)
    print("Recommendations for User 1:")
    print(recommendations[['name', 'price', 'rating']])

Why we do this: This is the heart of our AI shopping system. The code filters products based on user preferences and ranks them by rating to provide personalized recommendations.

Step 5: Run the Complete System

Now, let's create a main script to run everything together. Create a file called main.py:

from products import create_sample_products
from users import create_sample_users
from recommender import ShoppingRecommender


def main():
    print("AI Shopping Recommendation System")
    print("==================================")
    
    # Load datasets
    products_df = create_sample_products()
    users_df = create_sample_users()
    
    # Create recommender
    recommender = ShoppingRecommender(products_df, users_df)
    
    # Show all products
    print("\nAll Available Products:")
    print(products_df[['name', 'price', 'rating']])
    
    # Get recommendations for each user
    for user_id in users_df['user_id']:
        print(f"\nRecommendations for User {user_id}:")
        recommendations = recommender.recommend_products(user_id)
        print(recommendations[['name', 'price', 'rating']])
        
    # Show similar products
    print("\nSimilar Products to Wireless Headphones:")
    similar = recommender.get_similar_products(1)
    print(similar[['name', 'price', 'rating']])

if __name__ == '__main__':
    main()

Why we do this: This script ties everything together, showing how our AI system would work in a real shopping scenario. It demonstrates both personalized recommendations and product similarity features.

Step 6: Test Your System

Run your complete system by executing the following command in your terminal:

python main.py

You should see output showing all products, personalized recommendations for each user, and similar products. This simulates how Amazon's AI chatbot might recommend products to users during Prime Day.

Summary

In this tutorial, you've built a simple AI-powered shopping recommendation system that mimics the technology behind Amazon's Prime Day success. You've learned how to:

Create product and user data structures
Build a recommendation engine that filters products based on user preferences
Implement a basic similarity algorithm to find related products
Run a complete shopping recommendation system

This system demonstrates the core concepts that make AI shopping effective - understanding user preferences and matching them with relevant products. While this is a simplified version, real systems like Amazon's would use more complex algorithms, larger datasets, and machine learning models to provide even better recommendations.