Portfolio & Case Studies

Selected examples of completed analytics, machine learning, and software projects demonstrating production-ready thinking, disciplined engineering practices, and decision-focused design.

EMS Demand & Operational Analytics

Challenge

Emergency services operate under fluctuating demand patterns that make staffing and coverage decisions difficult without data-driven insight.

Approach

Analyzed EMS call and operational data to identify demand patterns and workload drivers
Applied statistical and predictive modeling techniques to understand trends over time
Structured analytical outputs to support operational decision-making rather than academic experimentation

Outcome

Improved visibility into EMS call behavior and operational drivers
Established a reusable analytical framework for future forecasting and staffing studies
Demonstrated how data science can support real-world public safety decisions

Capabilities demonstrated: Predictive analytics, operational data analysis, public safety modeling

Secure Event Management & User Systems

Challenge

Organizations require secure, role-based systems to manage events, users, and permissions while maintaining auditability and reliability.

Approach

Designed and implemented a RESTful backend system using modern architectural patterns
Integrated role-based access control and authentication workflows
Implemented validation, security controls, and automated testing to ensure system integrity

Outcome

Delivered a fully functional, secure backend system
Demonstrated production-ready coding standards and testing discipline
Showcased real-world software engineering practices beyond prototype-level development

Capabilities demonstrated: API design, authentication and authorization, secure systems, automated testing

Machine Learning Model Evaluation & Comparison

Challenge

Selecting the appropriate machine learning model requires objective comparison across multiple performance metrics rather than reliance on accuracy alone.

Approach

Implemented and evaluated multiple classification algorithms
Applied cross-validation to ensure generalizable performance
Compared models using confusion matrices and performance metrics
Documented results in a structured, reproducible format

Outcome

Clear comparison of classification models and their tradeoffs
Demonstrated disciplined evaluation methodology aligned with real-world ML workflows
Reinforced best practices in model validation and interpretation

Capabilities demonstrated: Supervised learning, model evaluation, cross-validation, performance metrics

Explainable Machine Learning for Medical Classification

Challenge

Healthcare-related classification problems require transparent and interpretable models that support understanding and accountability.

Approach

Built decision tree models designed for interpretability
Applied entropy-based splitting methods
Evaluated model performance using confusion matrices and classification metrics

Outcome

Delivered interpretable machine learning models suitable for healthcare contexts
Demonstrated explainable AI techniques appropriate for regulated environments

Capabilities demonstrated: Decision trees, explainable machine learning, healthcare analytics

Environmental Risk Analytics & Visualization

Challenge

Environmental risk data is often complex and difficult to communicate effectively to non-technical stakeholders.

Approach

Structured environmental and risk-related data into a clear analytical framework
Designed a lightweight web-based interface for communicating insights
Focused on usability, clarity, and stakeholder comprehension

Outcome

Improved accessibility of environmental risk insights
Demonstrated the ability to translate analytical work into stakeholder-facing tools

Capabilities demonstrated: Environmental analytics, data visualization, web-based decision support

Analytics Engineering & Data Science Foundations

Focus Areas

Regression modeling and statistical analysis
Association rule mining and pattern discovery
Clean code practices, testing, and reproducibility
Translating theoretical concepts into working analytical systems

Outcome

Strong foundation across core analytics and machine learning techniques
Demonstrated consistency in engineering discipline and analytical rigor
Built reusable patterns applicable to applied analytics and AI projects

What These Case Studies Demonstrate

Across these completed projects, Mullen Analytics consistently delivers:

Decision-focused analytics, not academic exercises
Secure, testable, and maintainable systems
Interpretable and defensible machine learning models
Applied analytics across public safety, healthcare, and environmental contexts

Explore Further

Request a capabilities briefingSchedule a strategy consultation