YICEN LIU

Automated Invoicing System for Property Services https://github.khoury.northeastern.edu/yicenliu/Automated_Invoicing_System.git

• Designed and developed an invoicing system for a property services company that specializes in various homeowner services in Java
• Developed pricing algorithms to calculate service costs based on a variety of factors to adapt to a broad spectrum of services from interior to exterior, ensuring that the invoicing system could accommodate a diverse range of home improvement tasks
• Employed UML diagrams for clear articulation of the system's architecture and used Domain-Driven Design (DDD) to define the interactions between different subsystems
• Developed unit and integration tests following the Test-Driven Development approach, utilizing frameworks like JUnit and JaCoCo

Vehicle Registration System https://github.com/vivianliu666/Vehicle-Registration-System.git

• Engineered a Java-based Vehicle Registration System using MVC architecture, utilizing subtype polymorphism and generics for flexible data management
• Applied Domain-Driven Design(DDD) to define key domains—Violation, DriverLicense, VehicleHistory—as bounded contexts
• Implemented singleton and composite design patterns to optimize component management, and achieved low coupling via stamp coupling
• Validated using nested inner classes and Test-Driven Development(TDD) with JUnit and JaCoCo for code coverage analysis

Development of Custom Encryption Program

• Built a Micro-service system in Python; Designed and implemented a series of fundamental encryption algorithms with a user-friendly interface to function in encoding and decoding messages to maintain the user's information privacy.
• Developed distinct modules for each encryption algorithm, with each module executing specific cipher techniques. Additional helper functions were also integrated as necessary to optimize the process.

Quantitative analysis - Social Analysis of International Large-Scale Assessments

• Explored datasets, designed and implemented Random Forest, XGBoost, Ridge Regression models with Python Seaborn and Stata to predict model assumptions in different countries
• Utilize pre-processing methods including Standard Scaler, Min-Max Scaler, ordinal encoding, random undersampling, random oversampling, ensemble resampling to practice data feature dimension reduction
• Implemented Linear Regression model, Logistic Regression model, and Neural Network model using Python
• Used grid search to tune hyperparameters; Presented at conference

Education sector analysis of Uganda

• Collected data from raw sources website; Cleaned data, executed comprehensive research and data synthesis on Uganda's education system. Analyzed how teacher-student ratio, teaching quality, and educational budget variation influence overall academic performance and achievement using R and STATA
• Used different graphs to visually underscore the system's prevailing challenges including scatter plots, correlation maps, stacked bar plots, facet graphs, and line graphs