Terra Store Recommendation System

How To Run

# Build Docker Container
docker build -t terra_app .

# Running Docker Container
docker run -p 8501:8501 terra_app

Project Documentation

Dataset Strategy

Customer Interactions:
1. Customer ID
2. Page views
3. Time spent on the website
Purchase History:
1. Customer ID
2. Product ID
3. Purchase date
Product Details:
1. Product ID
2. Category
3. Price
4. Ratings

Due lack of provided dataset, we using syntentic generator libraries called sdv to generate n-sample of rows according datasets.

Data Analytics

Customer has their own similar pattern about product purchases, this pattern might be generate because not much variety from sytentic data generator.

Model Choosen

We choose K-Nearest Neigbours in terms of Content Based Recommendation System due lack of variety dataset and limited scope of products.

Evaluation Metrics

We choose average precision, recall, and f1-score at 5-shot sampling recommendation data, with result all of them is 100% due dataset doesnt have enough variety.