Week 12 : Loaded Final Features & Recommendations

1. Introduction

In Week 12, our focus is on **loading the final features** and generating **product recommendations** for users using collaborative filtering. We build the **User-Item matrix**, compute **Item Similarity**, predict user ratings, and extract the **top 3 recommended products per user**. This ensures our recommendation model is ready for deployment and evaluation.

2. Objectives

Load final feature dataset from Week 11 outputs.
Construct the User-Item matrix representing user-product interactions.
Compute Item Similarity Matrix using collaborative filtering techniques.
Predict product ratings for each user.
Generate and save top 3 product recommendations per user.
Document key metrics and milestones for Week 12.

3. Week 12 Tasks / Assignment

Load the final features dataset prepared in Week 11.
Create User-Item matrix with users as rows and products as columns.
Compute Item Similarity Matrix using cosine similarity.
Predict ratings for all users based on item similarity.
Extract top 3 recommended products per user and save as 'week12_recommendations.csv'.
Visualize sample outputs for verification.

4. Outputs

4.1 Final Features (Sample)

Unique id	channel_name	category	Sub-category	Customer Remarks	Order_id	order_date_time	Issue_reported at	issue_responded	Survey_response_Date	Customer_City	Product_category	Item_price	connected_handling_time	Agent_name	Supervisor	Manager	Tenure Bucket	Agent Shift	CSAT Score
7e9ae164-6a8b-4521-a2d4-58f7c9fff13f	Outcall	Product Queries	Life Insurance	nan	c27c9bb4-fa36-4140-9f1f-21009254ffdb	nan	01/08/2023 11:13	01/08/2023 11:47	01-Aug-23	nan	nan	nan	nan	Richard Buchanan	Mason Gupta	Jennifer Nguyen	On Job Training	Morning	5
b07ec1b0-f376-43b6-86df-ec03da3b2e16	Outcall	Product Queries	Product Specific Information	nan	d406b0c7-ce17-4654-b9de-f08d421254bd	nan	01/08/2023 12:52	01/08/2023 12:54	01-Aug-23	nan	nan	nan	nan	Vicki Collins	Dylan Kim	Michael Lee	>90	Morning	5

4.2 User-Item Matrix Sample

Unique id	Affiliates	Books & General merchandise	Electronics	Furniture	GiftCard	Home	Home Appliences	LifeStyle	Mobile
0001f06a-bd9a-4888-9cf9-0a094a15eaf2	0	5	0	0	0	0	0	0	0
00053812-fb74-4294-a9d1-5167c5a36c7d	0	0	0	0	0	0	0	1	0

4.3 Item Similarity Matrix Sample

Product_category	Affiliates	Books & General merchandise	Electronics	Furniture	GiftCard	Home	Home Appliences	LifeStyle	Mobile
Affiliates	1	0	0	0	0	0	0	0	0
Books & General merchandise	0	1	0	0	0	0	0	0	0

4.4 Predicted Ratings Sample

Unique id	Affiliates	Books & General merchandise	Electronics	Furniture	GiftCard	Home	Home Appliences	LifeStyle	Mobile
0001f06a-bd9a-4888-9cf9-0a094a15eaf2	0	5	0	0	0	0	0	0	0
00053812-fb74-4294-a9d1-5167c5a36c7d	0	0	0	0	0	0	0	1	0

4.5 Top 3 Recommended Products per User

UserID	Top_Products
0001f06a-bd9a-4888-9cf9-0a094a15eaf2	['Books & General merchandise', 'Affiliates', 'Electronics']
00053812-fb74-4294-a9d1-5167c5a36c7d	['LifeStyle', 'Affiliates', 'Books & General merchandise']

5. Key Metrics

Total users processed: 50,000+
Total products: 10
User-Item matrix size: 50,000 x 10
Top 3 product recommendations generated per user
Recommendations saved successfully to 'week12_recommendations.csv'

6. Week 12 Milestone

Loaded final features from Week 11 outputs
Built User-Item matrix for collaborative filtering
Calculated Item Similarity Matrix
Generated predicted ratings for all users
Extracted and saved top 3 recommended products per user
Completed Week 12 assignment successfully