A collection of data science and machine learning projects I developed during my internship at SocialPrachar, Hyderabad. These projects focus on diverse domains, including employee attrition analysis, face detection, and product recommendation, with a goal to showcase various techniques in machine learning and data processing.
- Description: This project aims to predict employee attrition using machine learning techniques. It starts with an in-depth Exploratory Data Analysis (EDA) to understand the data and extract insights about the factors affecting employee turnover.
- Technologies Used: The project employs a Random Forest Classifier to predict whether an employee will leave or stay, based on factors like age, job role, satisfaction level, etc.
- Skills Gained: Hands-on experience with data preprocessing, feature selection, and applying classification algorithms to solve real-world HR challenges.
- Description: This project demonstrates face detection and recognition using the FaceNet model, a state-of-the-art deep learning architecture for facial recognition and feature extraction. The primary goal is to detect faces in images and extract embeddings for further identification or verification purposes.
- Technologies Used:
- FaceNet Model: Utilized for high-accuracy face detection, feature extraction, and facial recognition tasks.
- Google Sheets API: Implemented to store security logs, which can be used to maintain records of detected individuals for security purposes.
- Skills Gained: Working with neural networks for computer vision tasks, using APIs for data logging, and understanding face embeddings for recognition.
- Description: This project focuses on building a Product Recommendation System that suggests products to users based on their previous purchases and interests. It uses natural language processing techniques to generate product embeddings, helping recommend similar items.
- Technologies Used:
- Doc2Vec: A variant of Word2Vec used for creating vector representations of product descriptions. This helps the model understand the relationships between different products and make recommendations.
- Skills Gained: Experience with recommendation systems, applying Doc2Vec to create vector representations, and building a content-based recommendation system for e-commerce.
- Description: This project is an integration of the previous projects into a comprehensive system that provides personalized product recommendations to customers based on face detection and previous purchase history.
- It begins with using the Face Detection Model (FaceNet) to identify customers in real-time.
- Once a customer is identified, the Product Recommendation System is used to recommend products tailored to their preferences and purchase history.
- Google Firebase is used to securely store customer data, including personal details and purchase logs.
- Technologies Used:
- FaceNet for real-time face recognition of customers.
- Google Firebase for storing customer data and managing secure access.
- Product Recommendation System from Project 3 for suggesting products to the detected customers.
- Skills Gained: Combining multiple machine learning models, using cloud-based databases for data storage, and delivering an end-to-end product recommendation solution.
- Developed a full-stack machine learning solution that integrates multiple models and techniques for real-world applications.
- Enhanced understanding of data preprocessing, EDA, and model selection.
- Gained expertise in deploying cloud-based services like Google Sheets API and Google Firebase for data management.
- Applied Doc2Vec for product recommendation, and successfully leveraged deep learning models like FaceNet for face detection and recognition.