Preethi B

Using data to verify similarities is the secret to how Amazon, Flipkart, or YouTube suggest alternatives. The recommendation system is a similarity-based modeling technique that verifies the association of data points and basket preferences based on the same.

Outcomes include understanding association-based models, including the patterns of commonly used recommenders encountered daily, and building basic recommenders that suit business needs.

The aim of this project is to perform a comprehensive exploratory data analysis (EDA) on the datasets to understand their underlying structure, identify patterns, detect anomalies, and summarize the main characteristics using both visual and quantitative techniques. The analysis will cover univariate, bivariate, and multivariate analysis to gain deeper insights into the data.

KEY FEATURES
• Univariate, bivariate, and multivariate analysis
• Categorical variable encoding
• Normalization and scaling
• Missing value handling
• Data visualization and storytelling

This project showcases the complete lifecycle of a machine learning model from development to deployment and productionalization.

I serialized the model using Python's pickle library, allowing efficient model storage and reuse.

To facilitate accessibility, I implemented an API using Flask, creating endpoints that handle JSON inputs and outputs for smooth integration. I containerized the Flask application using Docker for development and production by writing and optimizing Dockerfiles to create efficient Docker images.

Additionally, I orchestrated the deployment using Kubernetes, configuring deployments and services for scalability, load balancing, and rolling updates. I integrated monitoring and logging solutions to track the application's performance and quickly address any issues.