My Projects
Checkout the projects I have worked on in my past.
Predictive Modeling for Bitcoin Price Dynamics Using Time-Series Data (LSTM)
Keywords: Python, LSTM Model, Big Data Analysis
During my Big Data coursework, I ventured into the realm of financial forecasting, specifically aiming to predict the dynamic price movements of Bitcoin. Leveraging the power of Long Short-Term Memory (LSTM) networks, a type of Recurrent Neural Network (RNN) adept at handling sequential data, I embarked on a project to dissect and model temporal patterns hidden within Bitcoin's historical price data. This endeavor involved meticulously constructing time-series features, carefully selecting hyperparameters, and rigorously calibrating the LSTM model through numerous training epochs. To gauge the model's effectiveness, I employed diverse metrics, delving beyond basic accuracy measures to unearth deeper insights into its strengths and weaknesses. This project fostered a profound understanding of both LSTM networks and the complexities of cryptocurrency price dynamics, serving as a springboard for further exploration into the fascinating world of financial prediction.
Pottery Studio
Keywords: Front end, REACT JS, JavaScript, MongoDB
In my web programming class, I took the lead in developing a dynamic e-commerce website for "Refined Pottery Studio," a hypothetical business selling stunning handcrafted ceramics and pottery classes. Collaborating within a 5-member team, I utilized my skills of front-end technologies like Express JS, React, HTML, and MongoDB to create a user-friendly and visually appealing online experience. Our project garnered an impressive A grade, reflecting the seamless integration of class registration, product browsing, and a secure checkout process. This project not only honed my technical skills but also demonstrated my leadership capabilities in driving a team towards collective success.
Predictive Text Generator
Keywords: n-gram, NLP, Python, RNN
During my NLP coursework, I journeyed into the fascinating world of language prediction by building a unigram and bigram-based language model. Using Python and its libraries, I constructed the model from scratch, feeding it a corpus of Chicago hotel reviews meticulously divided into training and validation sets. By evaluating the perplexity metric on the validation data, I assessed the model's ability to predict the next word in a sequence, gaining valuable insights into the strengths and limitations of each approach. This project not only solidified my understanding of statistical language models but also equipped me with practical skills in Python and NLP methodologies.
Sentiment analysis on Yelp reviews
Keywords: NLP, Python, FFNN, RNN
In my NLP class, I tackled the challenge of understanding customer sentiment in Yelp reviews. I evaluated two distinct approaches: a Feedforward Neural Network (FFNN) utilizing fixed word vectors and Softmax for predictions, and a Recurrent Neural Network (RNN) leveraging sequential data, word embeddings, and summed hidden states for output. Through meticulous hyperparameter tuning of hidden sizes and training epochs, I optimized both models. Analyzing accuracy and delving deeper in detailed reports, I conducted a comparative analysis to uncover each model's strengths and weaknesses. This project not only sharpened my NLP skills but also fostered a critical understanding of model selection and evaluation for real-world sentiment analysis tasks.
Library Catalogue
Keywords: CSS, Python Flask, JWT
Driven by a passion for books and a desire to optimize my reading experience, I embarked on a personal project to create a web-based library catalogue system. Utilizing Python-Flask for the backend, HTML and CSS for a user-friendly interface, and JWT for secure authentication, I brought this vision to life. The system empowers me to effortlessly add, edit, and delete book information, keeping my library meticulously organized. To spark spontaneous reading adventures, I even integrated a "random next read" generator, adding a touch of serendipity to my literary journey. Deploying the application on Heroku made it accessible from anywhere, showcasing my ability to bring projects from conception to real-world implementation.
Predicting personality using Social Data
Keywords: Skylearn, Pandas, Python Libraries
During my undergraduate studies, I explored the fascinating intersection of psychology and machine learning by tackling the challenge of predicting personality traits from social media data. Utilizing the OCEAN model (Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism), I leveraged the power of scikit-learn and Python to analyze user data and derive insights into their underlying personalities. This project involved rigorous data preprocessing, feature engineering, and evaluation of various machine learning algorithms. Through this exploration, I gained a deeper understanding of the relationship between online behavior and personality, honing my data science skills and contributing to the growing field of computational personality prediction.
