EmotiScan brings the magic of machine learning to recognize emotions and expressions in pictures. With EmotiScan, you can upload an image and receive predictions about the emotions and facial action units displayed in the photo. It's an enjoyable way to explore the subtleties of facial expressions and understand the underlying emotions.

QueryPDF is a web application designed to enhance document analysis by intelligently retrieving relevant sentences based on user queries. Traditional search methods often miss nuanced information in documents, requiring exact keyword matches. QueryPDF overcomes this limitation using advanced sentence transformer models to identify contextually relevant sentences, even when specific keywords are absent.

This project focuses on building a robust deep learning model for bird species classification using TensorFlow and the EfficientNetB0 architecture. The dataset consists of a diverse collection of bird images, and the goal is to train a model that accurately identifies and classifies different bird species.

Our primary objective was to leverage data-driven techniques to enhance the explainability of machine learning models in the context of finance. We focused on creating a prediction model to evaluate risk and adapted it to provide not only binary predictions but also a general confidence level, offering valuable insights for financial decision-makers.

The Death Rate Tracker (2019) Shiny App is an interactive data visualization tool built using Shiny, Leaflet, Plotly, and other R packages. It allows users to explore and analyze death rates across different countries, age groups, and causes of death for the year 2019.

This project visualizes the global connection between plastic waste and Gross Domestic Product (GDP), aiming to raise awareness among governments, waste management officials, and the public. This project originated from the inspiration of a visualization found on Behance. While recognizing the original visualization's merit, I identified some areas for improvement to better align with the visualisation's objective.