My Projects

A collection of applications, tools, and experiments I’ve built across data engineering, web development, and analytics.

Keyword Bot: An AI YouTube Video Keyword Generator

In Development

A smart, AI keyword generator that reads your YouTube video title, compares it to a dataset of real videos, and builds a pool of highly relevant tags. It currently uses custom NLP heuristics for relevance scoring, with a Genetic Algorithm coming next to polish everything into a perfect 500-character keyword list.

Tech: Python 3, NumPy, SpaCy, Genetic Algorithms

View on GitHub


# ===============================================
# 1. TITLE SEMANTIC SIMILARITY (weak weight)
# ===============================================
row_tokens = self.tp.tokenize(entry.title)
row_vec = self.tp.vectorize(row_tokens)
title_sim = self.tp.similarity(input_vec, row_vec)
      if title_sim < 0:
          title_sim = 0.0

Keyword Bot: An AI YouTube Video Keyword Generator

ThumbsUp - An Apple Watch 'Likes' System

In Development

An Apple Watch app prototype inspired by the 'likes' system in Death Stranding. This app allows users to send 'likes' to one another over local AirDrop connections, triggered by natural hand and wrist movements resembling a thumbs-up gesture. The goal is to bring a playful, gamified way of expressing gratitude to everyday interactions.

Tech: Swift, WatchOS, Motion Sensors, Haptics

View on GitHub


// Motion detection logic (SwiftUI + CoreMotion)
motionManager.startDeviceMotionUpdates(to: .main) { motion, error in
  if let attitude = motion?.attitude {
    if attitude.pitch > 1.0 { 
      sendLike() // thumbs up detected
    }
  }
}

ThumbsUp - An Apple Watch 'Likes' System

Office Space Rent Predictor

Completed

A Shiny web app built using R to predict commercial lease prices based on building size, industry type, move type, and location. Uses real data from the 2025 Elon Data Nexus DataFest competition.

Tech: R, Shiny, Data Visualization, Predictive Modeling

View on GitHub Live Demo


library(shiny)
ui <- fluidPage(
  titlePanel("Lease Price Predictor"),
  sidebarLayout(
    sidebarPanel(
      sliderInput("sqft", "Building Size (sq ft):", 500, 10000, 2000),
      selectInput("industry", "Industry Type:", c("Tech","Healthcare","Retail"))
    ),
    mainPanel(
      plotOutput("predictionPlot")
    )
  )
)

NBA Analytics - Indiana Pacers Assist Efficiency

Completed

A Google Colab project and research paper analyzing the assist distribution of the 2023–24 Indiana Pacers, before and after the Pascal Siakam trade. Using gramian matrices, I quantified passing relationships to measure changes in offensive flow, identified which players became the most common assist targets, and highlighted how roster moves impacted team efficiency.

Tech: Python, Google Colab, Pandas, Matplotlib, Seaborn

View on GitHub Read Paper


import pandas as pd
import numpy as np

# Load dataset
df = pd.read_csv("pacers_assists.csv")

# Build gramian matrix: passer (rows) × receiver (cols)
players = sorted(df["passer"].unique())
gram_matrix = pd.DataFrame(
    0, index=players, columns=players
)

for _, row in df.iterrows():
    gram_matrix.loc[row["passer"], row["receiver"]] += 1

print(gram_matrix.head())

NBA Analytics - Indiana Pacers Assist Efficiency