Credit Default Prediction using XGBoost
Developed a machine learning model to predict credit default risk using financial datasets. Applied SMOTE to handle class imbalance, used RandomizedSearchCV for hyperparameter tuning, and evaluated performance using ROC-AUC, F1-score, precision, and recall. Technologies: Python, Scikit-learn, XGBoost, SHAP, LIME, SMOTE.
Exploratory Data Analysis (EDA)
Performed exploratory data analysis to understand patterns and trends, cleaned data by handling missing values and detecting outliers, and used histograms, box plots, and heatmaps to analyze distributions and relationships. Technologies: Python, Pandas, NumPy, Matplotlib, Seaborn, Jupyter Notebook.
Online Bookstore Web Application
Built a full-stack web application for managing online book purchases and records, developed a responsive frontend, implemented backend functionality using Django, and managed book records in a database. Technologies: HTML, CSS, JavaScript, Python, Django, Bootstrap.
Medicine and Science Information Website
Developed a responsive React.js platform providing information about medicines, diseases, and basic health science topics. Implemented search and category features and focused on clean, accessible content. Technologies: HTML, CSS, JavaScript, React.js.
Invoice Generator application Website
Developed a web-based invoice generator for creating professional invoices and downloading them as PDFs. Implemented customer and product inputs, multiple billing items, automatic price calculation, dynamic totals, live invoice preview, and PDF generation using React PDF. Technologies: HTML, CSS, JavaScript, React.js, @react-pdf/renderer.