Product Recommendation(Walmart data)

Project Overview:
This project focuses on building a product recommendation system for a grocery dataset using both popularity-based and collaborative filtering techniques. The system suggests products to users based on their preferences and past behaviors, enhancing the shopping experience.

Features:
PRODUCT_NAME: Name of the product.
CATEGORY: Product category.
PRICE_RETAIL: Retail price of the product.
PRODUCT_URL: URL of the product for purchase.
PRODUCT_IMAGE: Image URL for the product display.

Process and Key Steps:

Exploratory Data Analysis (EDA):
Checked for missing values and duplicates.
Displayed key dataset characteristics to understand product categories and price ranges.

Recommendation Techniques:

Popularity-Based Recommendation System:
Selected top 5 popular products based on price (products with retail prices under $50).
Displayed relevant details, including product name, category, URL, and image for ease of access.

Collaborative Filtering Recommendation System:
Implemented using Cosine Similarity on a pivot table of products and categories.
Generated similarity scores between products and recommended products that are similar to a given product.

Conclusion:
Popularity-Based Recommendations: Simple and effective for recommending frequently purchased or popular items to all users.
Collaborative Filtering: More personalized recommendations based on user preferences, though it requires sufficient data on user behavior to perform optimally.

Recommendation for Future Work:
The model could be improved by incorporating user purchase history for a more robust recommendation system.
Introducing user behavior features such as rating or frequency of purchases can enhance the accuracy of collaborative filtering models.