Data Analyst | Software Developer
Python • SQL • Tableau

Data is the driving force for decision making and I enjoy diving head first into the data to find those hidden clues you would find else where.

Currently: Looking for fulltime work whilst creating personal projects.

📽️ Box Office Predictor and Analysis



Business Problem: People are not going to the movies anymore, Streaming is gradually taking over, and Ticket prices are slowly rising.

Analysis: Multi-decade analysis analyzing the relationship between ticket prices, box office turnout, movie genres, and cinema culture from 1977-Present.

Key Finding: TBA:

...TBA
...TBA
...TBA

Business Applications/Solution:

...TBA
...TBA
...TBA

Skills: Data Visualization • Regulatory Reporting • APIs

🚓 Violence Detection Using Ensemble Learning



Find my detailed report here!

Business Problem:
• When crimes are commited there is a large time gap between the crime and action taken by first responders. This time gap leads to criminals fleeing the scene and in some cases getting away from the crime, severe injury left unattended, less incentive to stop crimes, etc.

Motives:
• Current models, while powerful, often face challenges in handling the complexity and variability of real-world violent scenarios. • A single model no matter how sophisticated, could struggle to capture the full range of these variations due to its finite capacity and the specific biases it undergoes and develops while training. • To enhance and refine existing deep learning methods for violence detection.

Goal:
• We propose an ensemble method using the stacking technique, which combines the outputs of a 3D CNN, a simple CNN, and an RNN. The goal is to improve existing deep learning benchmarks that were trained on surveillance data by creating a more thorough violence detection method.

Key Findings:

By integrating multiple models, stacking can reduce overfitting and improve generalization, which is crucial for detecting 
violence due to the environmental nuances present in video frames. 
Overall, the ensemble method provides a significant improvement compared to the 3D CNN model alone, offering better 
generalization to unseen data—an essential feature when working with diverse datasets that may vary in characteristics 
such as location, population, and resolution.

Business Applications/Solution:

• Improvements in the area of surveillance, specifically in violence detection, is crucial for enhancing public safety and security in various environments in areas such as schools, public transportation, cities, and public areas.

• Effective surveillance systems can act as a deterrent to potential offenders, as the chances of detecting violent acts is increased.

• Authorities and first responders can be notified expeditiously and respond more swiftly to prevent harm and mitigate the escalation of violent incidents.

• Foster a sense of security within communities.

• Enhanced surveillance can provide valuable data for law enforcement and policymakers to understand the patterns and causes of violence, thus leading to more structured and better prepared industries surrounding violence prevention.

Skills: Data Visualization • Machine and Deep Learning

🌟 The Mall App



Overview

• Designed and maintained centralized dashboard systems to provide a friendly user interface for student activities, clubs, service times, and assignment monitoring. Implemented data storage techniques and database optimization using NoSQL systems to efficiently store and retrieve user data resulting in faster accessing times for in-app functions.

• Led comprehensive data collection, processing, and analysis for university’s digital platform serving 1,000+ users, implementing advanced analytics tracking using Firebase Analytics and Google Analytics allowing in timely distribution of app packages.

• Collaborated and with a cross-functional engineering team of 5 members to define critical data measurements, analyze system performance metrics, and develop effective in-app features for faculty and students resulting in a 25% in app traffic per school year.

• Presented technical findings and process improvement opportunities in bi-weekly meetings with faculty leadership to drive strategic decision-making and organizational performance evaluation resulting in effective and consistent shipment of new features.