Data Science Projects, 01/01/23, 01/01/24, Coventry, UK, Deep Learning and the Art of Recommendation: A Look at the Netflix Recommender System, Processed over 1M+ Netflix ratings using data normalization and sparse matrix embeddings, optimizing memory usage by 30% and improving data handling efficiency., Designed a hybrid recommendation model integrating collaborative filtering and deep learning, achieving a 16% increase in recommendation accuracy using TensorFlow, Keras, and Factorization Machines., Accelerated computational efficiency by 20% through parallel processing and batch optimization in Python., Engineered advanced features using Pandas and NumPy, incorporating time based interactions, user rating trends, and movie popularity, leading to a 14% improvement in model prediction accuracy., Analyzing Customer Data Tickets and Visualizing the Data using Tableau, Implemented a Logistic Regression model in Pyspark to predict customer's satisfaction, improving accuracy to 71% by optimizing feature selection and handling imbalanced classes., Performed extensive data preprocessing and feature engineering, reducing data inconsistencies by 20% through missing value imputation, categorical encoding, and normalization techniques., Developed interactive Tableau dashboards like Bar chart, Bubble chart, Heat map, Box plot, enabling a 25% faster identification of customer satisfaction trends based on gender, ticket type and priority supporting data-driven decision making.