Projects

Stock Movement Predictor

05/10/2021 Machine Learning Personal Project

Technology Stack:

I implemented a Neural Network with Sklearn in Python to predict if a stock will increase from historical trends. I leveraged the yFinance API to pull Adjusted-Closing market prices for stocks in the last 60 days as training data. Once the market prices were obtained, I normalized each stock price on a scale of 0 to 1 to represent the percent change from the previous day. I ran this model on thousands of Stock Tickers in Markets across the Globe. Click here to checkout the GitHub Repository for this project!

Wikipedia Search Engine Clone

04/18/2021 Web Development Course Project

Technology Stack:

I built a Web-Based search engine by processing and ranking over 3000 Wikipedia articles by writing Hadoop MapReduce programs. Each article was processed by removing stop words, removing punctuation, converting to lowercase, and computing the relative frequency of words within each article. Then, I computed TF-IDF (Term Frequency - Inverse Document Frequency) scores with given PageRank scores based off the given query in the Search box. This backend of the search engine was implemented in Python and I used Flask to design Front-End Web Pages by building 2 specific servers - 1 for indexing the articles and the other for querying the indexed database.

Time Customizable Abstractive Text Summarization

04/20/2021 Machine Learning Course Project

Technology Stack:

I implemented a neural network that aims to understand human emotion, sarcasm, and context clues by predicting summaries of user-specified lengths on news summaries. The model was built using Keras (TensorFlow) with a Seq2Seq model with LSTM, Fully Connected, Attention Mechanisms, Time Distributed, and ReLU Layers. I used the NLTK library to tokenize the dataset, stem verbs, convert to lowercase, and perform all preprocessing.

Instagram Clone

03/18/2021 Web Development Course Project

Technology Stack:

I recreated Instagram in Python by building personalized user profiles by using Flask with Jinja Templating and React. I used JavaScript for web-specific features (infinite scroll, double tap, history) and used SQL to store user data such as Name, Username, Posts, Comments, Followers, and Following in Oracle Tables. To ensure that the username and passwords were securely stored, I used SALT SHA512 Encryption Algorithm when inserting into the database and always Sanitized inputs in Python to guard again SQL injections.

Hotel Check-in Application

05/24/2020 Web Development Personal Project

Technology Stack:

I designed a cross-platform application (Web, iOS, Android) for users to book/edit hotel reservations by using Flutter. This app allows people to login into their individual accounts and either book, delete, or modify a reservation. Additionally, the app allows users to analyze price changes and available rooms during a given date range at a specific hotel. I used Google Firebase as the database to securely store and authenticate users and built app components and widgets with Dart programming language to take advantage of Google's documentation on Asynchronous Programming. Checkout the GitHub Repository for this project here!

Weather Effect of Music Analyzer

04/27/2020 Web Development Course Project

Technology Stack:

This Python script calls the Spotify API, MetaWeather API, and Billboards API to gather data as JSON objects and store them in a sqlite3 database. I collected data from 2013 to 2020 and produced visualizations using Matplotlib to analyze what genre of music is most popular during different weather conditions. Checkout the GitHub Repository for this project here!

Deep Neural Network for Image Classification

04/02/2020 Machine Learning Course Project

Technology Stack:

I designed a deep neural network using PyTorch to classify images of Pasta, Soup, Salad, Sandwich, and Cake with unknown labels. I built a Convolution Neural Network (CNN) with 3 Pooling Layers, 3 Convolution layers, 3 Fully Connected Layers, and 8 ReLU layers. The CNN takes a standard RGB image as its input and outputs which type of image it was (pasta, soup, salad, sandwich, and cake).

Web Scraping

04/02/2020 Web Development Course Project

Technology Stack:

This Python Project aims to take extract out data as a JSON object from a website and take out relevant information. This project summarizes all the headlines, authors, and published dates for each article posted on a website as an outputted text file.

Natural Language Processing Text Analyzer

02/15/2020 Machine Learning Course Project

Technology Stack:

This Python Project tries to predict the sentiment of a particular 'Rate my Professor' post. For example, the post 'This was the best professor I've ever had!' indicates that the post is most likely a positive review. However, the post 'I hated this class and the grading was highly unfair!' indicates that the post is most likely negative review. I used the AUROC performance metric to quantify how well my model performed on a validation data set. Additionally, I used a OneVsRest Multiclass Classifier and the NLTK Python library to build my model.

Android Application Development

05/15/2019 Web Development Course Project

Technology Stack:

I released an Android App called "Scientific Calculator" on the Google Play Store to compute different mathematical operations. I used Android Studio to modify various screen activities and wrote the app in Java. I learned to use a Gradle build System to store my application data as an APK file. See here to download the app off the Google Play Store! Checkout the GitHub Repository for this project here!

Computer Vision Based Facial Recognition

04/23/2019 Machine Learning Course Project

Technology Stack:

I programmed a facial recognition program in MATLAB from open sources to apply linear algebra principles and construct an eigenface. The model was trained on a dataset of many known faces. Then, it computed the basis, eigenvectors, covariant matrices of the dataset to display human-like combinations of various faces. These features were used to identify whose face the program took as an input.

Microprocessor-Based Music Synthesizer

04/20/2019 Course Project

Technology Stack: Verilog and Assembly Language

I engineered a keyboard synthesizer to teach people how to play musical notes on a digital piano. I developed a functional microprocessor by designing digital circuits in a hardware-descriptive language, Verilog. In order to simultaneously play sound and display the piano on a computer monitor screen, I programmed nonblocking input/output device drivers in assembly language for speakers, VGA monitors, and a PS/2 Keyboard.

Machine Learning-Based Post Classifier

04/20/2019 Machine Learning Course Project

Technology Stack:

I produced an algorithm to automatically identify the subject of a post from an online question and answer web service (Piazza). The program utilized introductory natural language processing, machine learning, and artificial intelligent techniques to compute log-probability score to classify each post. The statistical computations were made using a simplistic version of a Multi-Variate Bernoulli Naive Bayes Classifier.

Content-Aware Image Resizing Algorithm

02/15/2019 Machine Learning Course Project

Technology Stack:

I produced an algorithm to resize an image by computing a cost and energy matrix and removing unnecessary noise. The seam-carving algorithm utilized a test-driven development approach to resize its dimensions by desired input size. The implementation focused on practicing pointers, I/O operations, and C-style programming.