Welcome to my Data Science Portfolio. This collection represents my hands-on experience with a variety of data science techniques and tools. Here, you'll find projects that include Optimization, GIS Dashboarding, Image Processing with Computer Vision, and the development of various machine learning models such as Classification, Regression, CNNs, and Random Forest.
Each project showcases a specific set of skills and challenges, reflecting my practical approach to solving problems using data. Whether it's creating effective visualizations, processing images, or building predictive models, this portfolio is a window into my journey in data science.
The various projects are below:
This project is about my work with the administration of Bokaro (District in Eastern India with Population~2.5 million) to implement data-driven IT solutions, an interactive GIS Dashboard to track 30k incoming migrants and prevent the spread of the virus.
Real time LLM Recommendation Engine
This recommendation engine helps users identify the most suitable Large Language Model (LLM) for their use-case by aggregating data from various sources such as the Hugging Face Leaderboard, Twitter, and GitHub. It utilizes semantic search combined with real-time analytics to provide accurate recommendations.
In this project, we utilize various image processing and machine learning algorithms to detect and recognize text in documents and images. Another application area explored was embossed text detction. This project was carried out by me as part of the Summer Internship at IIT Delhi in the NVM Research group.
In this project, we build two approaches using Convolutional Neural Networks and some conventional image processing techniques for identifying four popular hand gestures - Thumbs Up, Stop, Punch, and Peace.
In this project, we build a system that can effectively detect and recognize handwritten equations from their images and further provide their solution if it exists.
The purpose of this project is to detect tampering of PAN Card using Computer Vision. This project will help different organizations in detecting whether the ID provided to them by their employees or customers or anyone is original or not.
We build a Random Forest Classifier for predicting the presence or abscence of Heart Disease in patients. Each attribute potentially contributes information that could be used to determine heart disease risk.
A Logistic Regression based classification model to predict Credit Card Defaulting. A particularly useful feature is that all classifiers (and linear models) are called using the same API, so it is easy to test between different models (see the sklearn-intro notebook for examples).
In this project, we predict Life expectancy of different countries. We will explore how immunization factors, mortality factors, economic factors, social factors and other health related factors affect Life expectancy of a country.
We build a recommendation system where we specifically focus on geographic locations to render the top 5 restaurants based on the current location of the user.
Amazon Food Reviews Sentiment Analysis
Using Roberta transformer from Huggingface and VADER models to find the sentiment of reviews of food items on Amazon.