Repository containing portfolio of data science projects completed by me for academic, self learning, and hobby purposes Using Python and iPython notebooks.
The projects involve various deep learning tasks using popular frameworks like PyTorch, TensorFlow, and Keras. They cover image retrieval, transfer learning for image classification, CIFAR-10 classification, fashion MNIST classification, dogs vs. cats classification, and digit recognition. These projects demonstrate the efficacy of deep learning in solving diverse computer vision problems
The presented computer vision projects encompass various image processing and manipulation tasks using OpenCV and other image processing techniques. The projects include image quilting with min cut for seamless texture synthesis, traditional fingerprint classification using Histogram of Oriented Gradients (HOG), JPEG compression implementation for file size reduction, indexed image compression through color clustering, Gaussian noise removal using Gaussian filters, cartoonization by applying edge detection and spatial transformations, face cartoonization through Haar Cascade face detection, and the creation of hybrid images blending low-frequency and high-frequency components for visual illusions. These projects demonstrate the versatility and capabilities of computer vision techniques in different image processing scenarios.
The machine learning projects cover a wide range of applications and techniques. They include a Spotify Music Recommender System with clustering and classification, TalkingData AdTracking Fraud Detection with imbalanced data handling, Machine Learning Cloud Services implementing various services, a Playground for fundamental algorithms, Mobile Price Classification achieving high accuracy in price classification, and Apartment Rental Price Prediction in Germany using regression with acceptable error rates. These projects showcase the versatility and effectiveness of machine learning in solving real-world problems.
The presented Artificial Intelligence projects demonstrate the application of various AI algorithms to solve challenging problems. The Snake Cube Solver efficiently tackles the Snake Cube puzzle using the A* search algorithm to find optimal solutions. The Sudoku Solver utilizes the Simulated Annealing algorithm to solve Sudoku puzzles by iteratively improving candidate solutions. Lastly, the Vacuum Cleaner project employs the BFS algorithm to design an intelligent agent capable of cleaning multiple rooms effectively. These projects showcase the power of AI techniques in solving complex problems and optimizing solutions.
The Data Analysis and Visualization projects involve exploring various datasets to gain insights and interpret the data. The "COVID-19 Data Analysis" project focuses on global COVID-19 data, while the "Data Analysis on Airbnb NYC dataset" examines Airbnb data in New York City. The "Data Analysis on Louisville Crime Reports dataset" deals with crime reports from Louisville government. Each project includes exploratory data analysis, data cleaning, statistical tests, and data visualization, showcasing the importance of these techniques in understanding and extracting valuable information from different datasets.