Deep Learning Based Breast Cancer Prediction
The motivation behind the project is to use deep learning to assist radiologists in making faster and more accurate diagnoses of breast and lung cancers. The system leverages a hybrid CNN-VGG19 model, which was trained on the Breast Ultrasound Images Dataset. This dataset is comprised of images classified as benign, normal, or malignant. The images are preprocessed by resizing them to 128x128 and normalising their pixel values. The model's training incorporates techniques like early stopping and model checkpointing to optimise performance. The final model achieved a high accuracy of 97.3% and a sensitivity of 97.1% in detecting malignancies. The system can be integrated into clinical workflows, potentially reducing diagnosis time from 20 minutes to under 2 minutes per case.
Customer analysis dashboard using Tableau
This project focused on creating a customer analysis dashboard using Tableau. The dashboard's purpose is to help businesses understand their customers' behavior and preferences by visualizing key metrics like demographics, purchase history, and buying patterns. The project uses a "Marketing campaign.csv" dataset and follows a detailed, step-by-step process to build an interactive dashboard. This process includes connecting to the data, creating individual worksheets with different charts and maps, combining them into a dashboard, and adding interactive elements like filters. The final dashboard provides a comprehensive view of the customer data, helping to answer business questions related to top customers, product purchases, marketing campaign performance, and customer churn.