Forecasting Fury: How Data Science Predicts and Prepares for Hurricanes

Used historical NOAA data to predict hurricane strength with ARIMA, Random Forest, and Logistic Regression models. Focused on seasonal patterns, feature engineering, and real-world forecasting insights.

Sentiment Classification of Amazon Product Reviews Using LSTM, GRU, and GloVe Embeddings

Developed and compared deep learning models (LSTM, GRU, and LSTM with GloVe) to classify Amazon product reviews into positive, neutral, or negative sentiment. Achieved ~88% accuracy using pre-trained embeddings for enhanced semantic understanding.

Human Gesture Recognition Using Skeleton-Based Deep Learning Models (CTR-GCN, 2S-AGCN, Transformer)

Built and compared multiple deep learning architectures (CTR-GCN, 2S-AGCN, Transformer Encoder) to classify human gestures from skeletal motion data using the NTU RGB+D dataset. Achieved up to 77.8% test accuracy with adaptive graph modeling and attention-based learning.

Customer Churn Prediction for Banking

Performed churn analysis on a banking dataset using logistic regression, random forest, and neural networks; applied oversampling to balance classes, enhancing model accuracy and yielding actionable insights into customer churn behavior.

London Bike Sharing Dashboard

Programmatically gathered data using Python. Explored, assessed and manipulated the data using the Pandas library. Finally, created impactful visualizations on Tableau using dynamic parameters and calculated fields.

Are you getting a good deal on your Toyota? An intro to linear regressions.

In this project, we pre-processed our data to obtain 100 cars out of 45,456 having 10 features. Then we created 4 LR models, which we then analyzed and compared. Finally, we found the best model and used it to predict the price of a 2015 Toyota Camry with 120,000 miles and 6 cylinders.

How studying Aboard impacts students mental Health Pt 1

In Pt 1 of this project, I conducted numerous statistical tests (correlation, ANOVA, Regression) to understand how the different features of studying abroad impacted the mental health of University students in Japan. Insights were used in Pt2 where the data was then further processed & visualised.

Netflix in 2020

Imported Netflix's data from an online source from the year 2020. Created an interactive dashboard sharing key information to users such as release dates, durations, ratings, descriptions, top movies in a given genre, etc.

Coffee Sales Analytics - Interactive Dashboard

Created an end-to-end Excel project. Gathered coffee shop data from GitHub, and transformed it using Pivot tables, and Pivot Charts. Finally built a dynamic interactive coffee sales dashboard.

NYMC Student Demographics

Extracted data from The Integrated Postsecondary Education Data System (IPEDS) Completions, Data USA, and other public sources. Cleaned and organized this data via MS Excel. Created a dashboard to help understand New York Medical College's academic landscape and aid in decision-making.

Investigating Netflix Movies - DataCamp

This project was a part of the 'Data Science Associate in Python Course' on DataCamp where I perfromed exploratory analysis on netflix_data.csv. I wrote code to find out the the most frequent movie duration in the 1990s and the number of short action movies released in 1990s.