Tripti Pandey

Results-driven Data Processing Manager with 13 years of experience leading large-scale market research data operations, delivering high-quality analytics, and driving process efficiency. Adept at managing cross-functional teams, optimizing workflows, and ensuring data integrity for strategic decision-making. Recently certified in Applied Data Science & AI from E&ICT IIT Roorkee, with hands-on expertise in machine learning, predictive analytics, and data visualization. Passionate about bridging the gap between business needs and advanced analytics to enable impactful, data-driven strategies.

Post Graduate Certificate in Applied Data Science and AI, E&ICT IIT Roorkee, July2025

Diabetes Prediction Using Machine Learning
PG Certificate Program in Applied Data Science and AI, E&ICT IIT Roorkee 

• Developed an end-to-end classification pipeline to predict diabetes risk using the Pima Indians Diabetes Dataset.
• Performed EDA with correlation heatmaps, distribution plots, and boxplots to identify trends and outliers; applied IQR method to mitigate skewness while retaining all records.
• Implemented and compared multiple models — Logistic Regression, Support Vector Classifier (SVC), Random Forest, Decision Tree, KNN, XGBoost — evaluating performance on training (80%) and test (20%) splits.
• Achieved best generalization with SVC and Logistic Regression (balanced training & validation accuracy), while identifying mild overfitting in XGBoost and tree-based models.
• Suggested future improvements with larger, diverse datasets and advanced hyperparameter tuning for enhanced accuracy and generalizability.
• Tools & Libraries: Python, Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn, Jupyter Notebook.
• GitHub: Capstone_Project_Tripti_Aug25/Diabetes_Prediction_Tripti_Project_Jul25.ipynb at main · Trptpndy/Capstone_Project_Tripti_Aug25 · GitHub

• Developed a predictive model to forecast hourly bike rental demand using historical usage and weather data, applying machine learning algorithms
• Achieved high-precision predictions, improving the efficiency of bike-sharing system management, and identifying key factors like weather and time of day
• Built a robust, scalable model pipeline that enhanced prediction accuracy and reliability