Ä¢¹½´«Ã½

California's climate disclosure laws, Senate Bills 253 and 261, require approximately 2,600 to4,100 companies to report greenhouse gas (GHG) emissions beginning in August 2026, yet fewaccessible tools exist to guide organizations through this complex regulatory process. Thisproject addresses that gap by developing a web-based Software-as-a-Service (SaaS)application that automates GHG emissions calculations and compliance report generation inaccordance with California's requirements. The system implements the GHG Protocol CorporateStandard's dual Scope 2 reporting methodology both location-based and market-basedcalculations alongside Scope 1 direct emissions, drawing on authoritative emission factordatabases from the EPA and the California Air Resources Board (CARB). A multi-tenantarchitecture supports secure, concurrent use across organizations, while an intuitive guidedinterface enables non-technical users to complete their first compliance report within one hour.The application generates California SB 253-compliant PDF reports complete with emissionsbreakdowns, methodology documentation, and audit trails. By making rigorous, accurate climatecompliance accessible without relying on expensive proprietary platforms, this system lowersthe barrier to regulatory adherence for small and mid-sized California businesses facing animminent and consequential new legal mandate.

ABSTRACT
Organizations across industries continue to rely on paper-based documentation that slows operations, complicates audits, and increases security risks. Existing solutions address parts of this problem but often carry licensing costs that put them out of reach for many organizations. Formigo is a web-based form management system designed to streamline data collection while safeguarding organizational boundaries. The platform enables any manager (aka "admin") to design customizable form templates with features like drag-and-drop interfaces, supporting dynamic fields, scheduling rules, and automated reminders. Employees (aka "users") can complete, save, and submit forms from any browser, with built-in auto‑save minimizing data loss during interruptions. Security is embedded in every stage of the system architecture: server‑side validation ensures that all API requests are authenticated and restricted to their respective organizations. Completed forms proceed through a structured review workflow, creating a tamper‑evident audit trail that reinforces accountability and compliance. Formigo demonstrates how enterprise‑grade security can exist on the web with a seamless user experience at an affordable, reasonable price.

This project presents the design and construction of a mobile application which enhances the experience of golf by simplifying scorekeeping and game setup. The app allows users to select a course, a specific game, and track player scores through a simple interface. A future integration feature will be an image based import of a course scorecard to create minimal input from the user. The application minimizes input and customization for multiple types of games and betting formats. Constructed with a modern mobile application framework and integrated with a backend database which supports course selection and future integration to allow users to add courses to the database. The application aims to simplify golf scoring on the course while introducing a digital solution for casual and competitive players. 

Cybersecurity is something most systems deal with today, so having a basic understanding of how attacks work is important. In this project, I set up a virtual lab using Kali Linux as the attacker machine and Ubuntu Server as the defender to simulate real-world cybersecurity scenarios. The goal was to show how common attacks, such as port scanning and basic intrusion attempts, actually work and how they can be recognized within a controlled environment.
The lab environment was created using Oracle VirtualBox, where both machines were connected through a private virtual network to allow direct communication. Using tools available in Kali Linux, I performed network scans and simulated different attack scenarios, then observed how the Ubuntu system responded, especially when identifying open ports, active services, and potential vulnerabilities. I also looked into different types of attacks to determine which ones were realistic and appropriate to demonstrate within the scope of this project.
After setting everything up, I ran multiple test scenarios, mainly focusing on repeated network scans and connection attempts, to evaluate system behavior and consistency in results. These tests showed that even simple configurations can expose weaknesses if a system is not properly secured. Overall, this project helped me better understand both the attacking and defending sides of cybersecurity and gave me a stronger foundation to build on.

Music streaming services like Spotify track millions of user listening habits, but current tools offer limited ways to explore complete listening history beyond annual summaries. Understanding long-term musical preferences and how they evolve provides valuable insights into personal identity and cultural engagement.
The purpose of this project was to develop AudioMemory, a web-based application that transforms raw Spotify streaming data into interactive visual timelines, allowing users to explore their complete listening history and discover patterns in their music consumption habits.
The application was built using Flask (Python) for backend data processing and Chart.js for frontend visualizations. Users first request their complete listening history from Spotify, which provides exported streaming data as JSON files. Once received, users upload these files through a drag-and-drop interface. The system parses thousands of listening records, filters plays by duration, calculates aggregate statistics, and identifies temporal patterns including monthly trends, hourly listening habits, and day-of-week preferences. The data is then visualized through interactive charts displaying monthly activity timelines, top artists and tracks, and listening pattern heat maps.
Testing with a dataset of 62,753 listening records spanning 13 months demonstrated the application successfully processes large datasets in under 2 seconds. The system identified 2,466 unique artists, 30,026 valid plays totaling 1,491 hours, and revealed distinct listening patterns throughout different times of day and week. Unlike existing tools focused on annual summaries, AudioMemory provides users with a complete retrospective view of their musical journey, helping them rediscover forgotten favorites and understand how their relationship with music has changed.

In the ongoing search for Supersymmetric (SUSY) particles, this project focuses on identifying potential signatures of the Supersymmetric Top Quark, or stop quark (t~), produced in proton-proton collisions at the Large Hadron Collider (LHC) at CERN. SUSY provides a theoretical framework extending the Standard Model, offering both a path toward unifying the fundamental forces and a compelling dark matter candidate—the lightest supersymmetric particle (LSP). Earlier stages of this research developed a CMSSW-based analysis framework using FWLiteHistograms to process simulated CMS collision data through di-lepton decay channels—electron-electron (ee), muon-muon (μμ), and electron-muon (eμ). Preselection cuts on kinematic variables such as missing transverse energy (MET), invariant mass, and lepton isolation were optimized to separate SUSY stop signals from dominant background processes including tt¯, WZ, WW, and Drell–Yan events. Building on that foundation, the current phase transitioned the analysis to a NanoAOD-based workflow utilizing the Coffea library and Awkward Arrays in a Jupyter environment, enabling faster and more scalable event processing. New b-tagging algorithms, jet energy scale (JES) corrections, and MET resolution refinements were implemented, alongside comprehensive systematic studies. The team is now awaiting updated simulation samples to validate the refined selection criteria and to train a boosted decision tree (BDT) classifier for improved SUSY signal discrimination. This work contributes to the broader effort to uncover evidence for supersymmetry—an essential step toward unifying the fundamental forces and identifying the nature of dark matter.

Algorithmic trading now dominates modern financial markets, yet building a personal automated trading system has historically required institutional resources and expertise beyond the reach of most individuals. This project develops a machine learning–based stock trading system capable of making autonomous buy, sell, and hold decisions across a portfolio of large-cap equities. A fifteen-year dataset of historical price and volume data is used to calculate technical indicators, including momentum oscillators, trend-following signals, and volatility measures. These serve as inputs to a supervised classification model trained to predict next-day price direction. The model’s signals are then translated into trading actions through a portfolio tracking framework that maintains a running record of holdings, cash, and net returns. Strategy performance is evaluated on five years of out-of-sample data and compared against an equal-weight buy-and-hold benchmark using risk-adjusted metrics including cumulative return, Sharpe ratio, and maximum drawdown. The final market timing strategy achieved a 357.9% cumulative return with a risk-adjusted Sharpe ratio of 1.41, outperforming the benchmark’s Sharpe of 1.21 despite the inclusion of transaction costs. These results suggest that a data-driven approach to market regime detection can generate a meaningful and measurable edge over passive market participation.

Sexual dimorphism describes trait differences between males and females of the same species. In primates, body size dimorphism can help with male-male competition for females. Adult male squirrel monkeys (genus Saimiri) undergo a hormonally-induced “fattening” in their 8-week breeding season. Possibly, male fattening is a sexually-selected trait that allows fatter males to have higher access to mates. This field study examined whether the male fattening trait in Saimiri collinsi is linked to advantages acquired during male-male competition events. We hypothesized fatter males would gain more access to females due to their higher competitive abilities. We predicted that more robust males would participate in and win more contests. We collected behavioral data on two squirrel monkey social groups during three mating seasons (2023-2025; N=414 contact hours), in Eastern Amazonia, Brazil. Study groups averaged  46 individuals, 12 were adult males. Males were classified on a visual scale of 1 to 3 (1=smallest, 3=fattest). Focal animal samples were used to construct nearest neighbor budgets for different males. We recorded the initiator, recipient, winner of male-male contests, and copulation occurrences. Grade 3 males were responsible for 88% of the copulations (N=50). During male-male contests, females were within 3m 73% of the time (N=164 observations). Grade 3 males won 67% of competitive interactions, initiating 71%. Fatter males are more competitively dominant and gain more access to females, suggesting that the degree of fattening of male S. collinsi affects their competitive ability and mating success. Funded by NSF (Award no. BCS-2140666).

This study evaluated the thermodynamics of metal ions binding to chelating agents by isothermal titration calorimetry (ITC). Corn cellulose was mechanically processed as a model system for a low-cost and used as a sustainable bio-adsorbent for heavy metal ions (cadmium and nickel) in solution. EDTA was used at varying concentrations, along with differences in metals and metal concentrations, used as a reference chelating agent for comparison of thermodynamic characteristics. Including binding affinity (Kd), stoichiometry of molecular interactions (n), and binding enthalpy (ΔH) changes. Initial results indicated that a 0.05 M EDTA solution demonstrated lower K d values and near 1:1 value related to n with metals, especially nickel. Moreover, high exothermic ΔH values were revealed at higher metal concentrations (up to 25 ppm), indicating that heat is released upon successful binding. Future studies will focus on testing additional metals, pH effects on EDTA complexation potential, and investigating additional metal concentration ranges.

This project studies when a topological space is metrizable, culminating in the Urysohn Metrization Theorem with focus on the roles of regularity, Hausdorffness, and second countability in determining when a topology arises from a metric. Urysohn’s Lemma provides the key step by separating disjoint closed sets with continuous functions. Using a countable family of such functions, the space is embedded into a countable product of intervals, where a metric can be defined and pulled back to the original space. Examples are included to show why each hypothesis is necessary and to illustrate the distinction between metrizable and non-metrizable spaces.

This project presents First Spin, a mobile application designed to improve the music discovery experience through a personalized, daily format. While modern streaming platforms provide recommendation systems, users often encounter overwhelming amounts of content or repetitive suggestions, making it difficult to engage meaningfully with new music. First Spin addresses this problem by delivering a small, curated set of songs each day, encouraging more intentional listening and reducing decision fatigue. The application incorporates user listening preferences to guide recommendations while still introducing new and diverse content. In addition, each daily selection includes contextual information about the songs and artists, allowing users to better understand and connect with the music. By combining personalization with a limited, structured approach, this project aims to create a more focused and engaging music discovery experience, while supporting future expansion and deeper integration with external music platforms. The application was created to support music discovery and encourage users to return each day to explore new music.

This project focuses on the design and implementation of the puzzle game Tetris within the context of software engineering and game development. The goal is to recreate the core mechanics of the game while maintaining a clear and well-structured object oriented design. A systemic development process is followed, beginning with planning and system modeling using Unified Modeling Language (UML) diagrams to define class structure, object relationships, and system behavior. The implementation reflects essential gameplay features, such as piece movement and rotation, collision detection, line-clearing mechanics, and game state management within a fixed grid environment. In addition to the standard gameplay, multiple game modes are incorporated to extend functionality and demonstrate flexibility in design, including variations that emphasize scoring and dynamic challenges. The project highlights the importance of organization, abstraction, and event-driven design in building interactive systems. Overall, it serves as a practical application of core programming and software engineering principles in the development of a structured and interactive game.

This project focuses on designing a life-sim video game that combines farming, fishing, and social interaction with characters based on different mythologies. Many life-sim games, like Stardew Valley, focus on building relationships and managing daily tasks, but what makes my game different is the infusion of multiple mythologies. The goal of this project is to give players a relaxing escape from their everyday lives while also teaching them about different mythologies and their gods. In this game, players help restore a small town while growing crops, fishing, and forming relationships with both regular townspeople and gods such as Ra, Zeus, and Odin. These gods can interact with the player in different ways, like giving small rewards, reacting to offerings, or being a larger part of the town. This project demonstrates a simple way to expand the life-sim genre by mixing everyday activities with mythological storytelling. 

As my Capstone project, I developed "Cyborgan", a cross-platform software synthesizer.
Cyborgan takes some inspiration from electric organs, with nine drawbars used to control the volumes of nine wavetable oscillators pitched at various multiples of the base pitch of a note played. Unlike electric organs, each of these oscillators may have their volumes further modulated by a corresponding envelope generator; in addition, a tenth envelope generator modulates the synthesizer's overall output volume. Each envelope generator incorporates six parameters, consisting of Delay, Attack, Hold, Decay, Sustain, and Release, granting fine control over how an oscillator's volume changes over time. The independence of each of the ten envelopes allows for the creation of evolving timbres atypical of both electric organs and conventional synthesizers, such as fading different oscillators in and out.
Cyborgan is cross-platform, compatible with MIDI, and usable as a standalone program or as a VST3 or AU plugin for a digital audio workstation.

Neurolytics is a health informatics system designed to improve the documentation and analysis of neurological episodes, addressing the limitation of traditional symptom tracking methods that rely heavily on unstructured patient recall. Conventional approaches to capturing neurological events often fail to capture subtle or non-obvious events, such as focal-aware or absence seizures, reducing their value for longitudinal analysis and clinical interpretation. The purpose of this project is to develop a structured framework for consistently capturing and analyzing neurological episodes across sensory, cognitive, motor, and autonomic domains. The system transforms user-reported data into standardized representations, enabling more effective comparison and analysis over time. Pattern recognition techniques are applied to identify recurring symptom clusters and relationships between events, supporting more consistent interpretation of patient-reported data. A modular, layered architecture was implemented to separate data collection, processing, and analytical components, ensuring scalability and maintainability of the system. Preliminary results indicate that structured episode logging improves the consistency of patient-reported data and enables the detection of recurring symptom patterns that may not be evident through traditional tracking methods. These findings suggest that Neurolytics can enhance clinical interpretation and support more informed neurological assessment by providing a more reliable and analyzable representation of patient experiences.

This project presents the design and implementation of a QR-code-based attendance and access control system tailored for gym environments. The system addresses the limitations of traditional manual attendance tracking by providing a more efficient and accurate method for recording member participation. Through QR code scanning or name-based search, the application enables rapid check-in while automatically associating each entry with a specific class session and timestamp. The system incorporates attendance policy enforcement based on member type, ensuring that gym rules are applied consistently without requiring manual oversight. In addition to improving operational efficiency, the application includes administrative features such as attendance summaries and reporting tools, which support data-driven decision-making related to class scheduling, staffing, and overall gym management. Designed as a browser-based solution, the system emphasizes accessibility, scalability, and ease of deployment in real-world settings. This project demonstrates how integrating simple digital technologies into existing workflows can significantly enhance accuracy, efficiency, and organizational insight within small business operations.

Jsynth1 is a software synthesizer which aims to emulate traditional analog synthesizerinstruments via digital software. Jsynth1 contains many industry standard features which requirestrict event flow and multithreaded code. These features include note triggered oscillators,envelopes, polyphony, midi control, UI, patch storage, and postprocessing effects. The mainfocus of my research is to outline how I went about organizing and connecting all of thesedynamic parts to create a cohesive instrument. The goal is to create software such that a personwith little to no knowledge of computer science or synthesis can adeptly utilize the instrument ina way that seems obvious. The user should always be in the position of making creativedecisions rather than the user being overwhelmed or confused. As for the backend, a majorchallenge is managing data flow for asynchronous events. Each note played on the keyboard isdynamically assigned a voice which triggers the assigned envelope, and the voice also needs tobe notified when a given key is released, triggering the release part of the envelope. My hope isthis research provides useful information and guidance to others who are interested in creatingdigital instruments.

Many guitar applications exist today, all with different features such as tuners, metronomes,backing track libraries, and interactive video lessons with chord recognition software. All ofthese applications, however, require subscriptions to access most features. This capstoneattempts to use open-source libraries to create an all-in-one guitar application that contains manysuch features. The pitch detection library CREPE was used in the implementation of a tuner andwas also used in conjunction with Spotify’s BasicPitch library and my own model, created tomap a given audio clip to a fret on a specific string, to automatically generate tabs while the useris playing their guitar. Open-source sounds were used to create a working metronome, and thebacking track generator relies on a Google cloud model to generate loopable audio based on userselected keywords. This app, therefore, is a free alternative to the subscription based guitar appsand was designed with the everyday guitarist in mind.

Topology is the study of invariant-preserving transformations of open sets, and has applications in molecular biology for modeling the structure and behavior of enzymes. Topologists Ernst and Sumners[1] previously published a theorem describing DNA recombination products of the bacterial enzyme TN3 resolvase. The theorem used the Cyclic Surgery Theorem, properties of 2-fold branched covers, and tangle theory to predict the equivalence class of the resulting DNA product and predict the structure of rarer recombination products. This project explains the important connection between biochemistry and topology and gives a detailed overview of the published theorem for a general audience. We present a centralized list of definitions supported by intuitive examples to demonstrate how topological concepts are utilized in the theorem’s proofs without requiring previous topology knowledge.