Steam Game Data Tracking and Analytics
An automated data pipeline that continuously monitors newly released and recently updated Steam games, capturing metrics such as player counts, prices, and review scores. Data is stored in a PostgreSQL database. The project leverages Python, Node.js, PostgreSQL, the Steam API, and Cron for automation. View the source code on https://github.com/DStark1024/steam-data-tracker
The project also features an interactive Tableau dashboard for exploratory analysis, allowing users to filter games by genre and category, compare key metrics, and visualize historical trends in player engagement, pricing, and review scores through dynamic charts and dashboards https://public.tableau.com/app/profile/dmitrii6272/viz/Steam_17822723818190/Dashboard
PubMed ML/NN Analytics
A data pipeline and analytics system for PubMed medical publications that involve the application of Machine Learning and Neural Networks in biomedical research. It collects and processes research articles using the PubMed API, performs MeSH-based topic analysis, and builds time-series trends of publication activity. The project includes ARIMA-based forecasting, visualization of research growth, and a REST API for accessing analytics results. Interactive visualization of PubMed ML/NN publication trends available at https://dstark.space/pubmed/. View the source code on https://github.com/DStark1024/pubmed-data-analytics.git
Customer Churn Survival Analysis
A survival analysis and machine learning project for predicting customer churn using a simulated e-commerce dataset obained form Kaggle. Since the dataset is intentionally simulated and does not contain strong real-world behavioral patterns, the model predictive performance is limited. This makes the project a useful example of the limitations of ML methods when applied to noise-dominated or pattern-free datasets. Regardless of the simulated nature of the dataset, this project is intended as a demonstration of practical skills in applying analytical and machine learning techniques to customer behavior data. View the source code on https://github.com/DStark1024/simulated-churn-survival-analysis.git