I am Elena

Next.js Crypto Dashboard

A modern cryptocurrency dashboard built with Next.js, providing users with features like real-time market analytics, user authentication, and portfolio management; developed as a group project.

• Real-Time Cryptocurrency Analytics: Track top gainers, losers, and market trends with data from CoinGecko API.
• Weather Insights: Integrated weather data to enhance user experience.
• Responsive Design: Optimized for desktop and mobile devices.
• Protected Routes: Access control and user authentication implemented for secure navigation.

Tech Events Board

Full-stack event management platform using React and Node.js with the following features:

• User authentication, event creation, and filtering options
• Responsive and modern UI powered by Tailwind CSS
• Backend built with Express and MongoDB
• JWT used for secure user sessions
• Note: As this project is hosted on a free platform, it might require a refresh to load initially.

Edamam Angular App

A responsive Angular application that helps users explore meal suggestions, search for recipes, analyze nutrition, and submit feedback. Built with Angular and deployed on Vercel.

• Meal Suggestions: Explore curated meal ideas tailored for users.
• Recipe Search: Seamlessly search recipes using the Edamam API's NLP-powered capabilities.
• Nutrition Analysis: Analyze recipes and food items with detailed nutritional insights.
• Feedback Form: Submit feedback through a styled and responsive form.
• Responsive Design: Optimized for desktop and mobile views with modern styling.

Beauty Store

Developed a React-based beauty store application with dynamic animations using GSAP and seamless navigation with React Router.

Property Agency

Responsive, modern website for property agencies using HTML, CSS, and Bootstrap.

Utility Scripts

A versatile collection of Python and Shell scripts designed for data processing, API interaction, ETL pipelines, and automation tasks.

• Python: Includes scripts for data pipelines, API calls, and XML-to-JSON conversion.
• Shell: Utility scripts for system information, basic arithmetic, text ciphering, and more.

Tableau Toronto Crime vs Rent Analysis

A comprehensive data analysis group project exploring the relationship between crime rates and rental prices in Toronto. The project integrates insights from over 23,000 crime records and 5,000 rental listings, using Tableau to visualize and uncover key urban trends.

• Analyzed crime patterns and rental market dynamics across Toronto neighborhoods.
• Created interactive dashboards highlighting correlations between urban safety and housing affordability.
• Identified actionable insights for residents, policymakers, and researchers.
• Collaborated with a team to design visualizations, including maps, bar charts, and heatmaps.

Leveraging Graph Databases for Healthcare Analytics

This group project utilized Neo4j graph databases to analyze the FDA Adverse Event Reporting System (FAERS) dataset. The insights reveal relationships between drugs, adverse reactions, demographics, and outcomes, providing actionable intelligence for healthcare decision-making.

• Implemented graph modeling for 23,000+ cases, transforming complex tabular data into insightful relationships.
• Applied advanced Cypher queries and the Louvain algorithm to uncover high-risk drugs and reaction clusters.
• Built predictive pipelines with machine learning for link prediction and node classification, achieving high accuracy.
• Highlighted critical insights on drugs, manufacturers, and demographics, guiding regulatory actions and patient safety measures.

Human Activity Recognition Using Sensor Data

Developed deep learning models to classify human activities based on smartphone accelerometer and gyroscope data. Achieved high accuracy through innovative preprocessing and model optimization techniques.

• Built a Convolutional Neural Network (CNN) achieving 93.24% accuracy for classifying six activities, including walking and sitting.
• Explored PCA for dimensionality reduction, optimizing CNN performance while identifying its impact on temporal data for LSTM models.
• Analyzed misclassification patterns, revealing overlaps between similar activities like walking upstairs and downstairs.
• Provided actionable insights for feature engineering and future architectural improvements.

Object Detection Using Faster R-CNN

Developed and evaluated an object detection pipeline using PyTorch's Faster R-CNN model. The project focuses on bounding box prediction, Intersection over Union (IoU), and performance metrics like precision, recall, and F1 score.

• Trained and fine-tuned Faster R-CNN for detecting objects in images with COCO-pretrained weights.
• Achieved high IoU values for key predictions, showcasing the model's ability to locate objects accurately.
• Implemented metrics to evaluate model performance: Precision (1.0), Recall (1.0), and F1 Score (1.0).
• Utilized PyTorch and torchvision for efficient model integration and inference.

Object Detection Using Faster R-CNN

Developed and evaluated an object detection pipeline using PyTorch's Faster R-CNN model. The project focuses on bounding box prediction, Intersection over Union (IoU), and performance metrics like precision, recall, and F1 score.

• Trained and fine-tuned Faster R-CNN for detecting objects in images with COCO-pretrained weights.
• Achieved high IoU values for key predictions, showcasing the model's ability to locate objects accurately.
• Implemented metrics to evaluate model performance: Precision (1.0), Recall (1.0), and F1 Score (1.0).
• Utilized PyTorch and torchvision for efficient model integration and inference.

Sleep Health and Lifestyle Analysis Dashboard

An interactive Shiny dashboard group project analyzing the relationship between sleep patterns, lifestyle factors, and health outcomes. The analysis is based on the Sleep Health and Lifestyle dataset from Kaggle, utilizing R and machine learning techniques to uncover actionable insights.

• Explored correlations between physical activity, stress, and sleep quality through interactive visualizations.
• Developed machine learning models to predict BMI categories with an 83% accuracy rate.
• Implemented filters and side-by-side analysis for in-depth exploration of key health metrics.
• Created visualizations, including scatter plots, heatmaps, and box plots, to uncover trends.

Analyzing the Effectiveness of Vaccination Campaigns in Reducing COVID-19 Deaths

This group project evaluates the global impact of COVID-19 vaccination campaigns on reducing death rates. Using the Our World in Data dataset, the project analyzes vaccination trends, performs machine learning predictions, and uncovers insights for public health strategies.

• Conducted data cleaning and exploratory data analysis on COVID-19 cases, deaths, and vaccination rates.
• Applied machine learning models (Linear Regression, Decision Tree Regressor, Random Forest, K Neighbors Regressor) to predict death-to-case ratios.
• Visualized vaccination trends and their correlation with reduced mortality through advanced visualizations.
• Highlighted actionable insights for public health policies to mitigate the impact of the pandemic.

Data Science & Machine Learning

Exploring the world of data through machine learning and analytics, driven by a passion for discovering insights and solving complex problems. Check more of my Projects on GitHub.