Overview

DynastyLab AI solved a $70+ game's biggest UX problem: players were making terrible upgrade decisions because they couldn't see the math behind 200+ upgrade paths. By building a custom efficiency scoring model and wrapping it in an intuitive dashboard, I created a tool that improved player decision making while demonstrating core BI and product management principles.

—————————

The Problem

"What should I Upgrade?"

Dynasty Mode forces players to allocate Skill Points across 1,000+ upgrade paths with zero decision support. Players see options like “5 SP for +3 Speed” or “12 SP for Silver Strong Grip,” yet the game offers no way to compare value, evaluate efficiency, or plan a progression strategy. This leads to wasted SP, inefficient builds, and frustration with high-cost abilities that offer unclear value.

—————————

Technical Approach

Data Architecture

I built a SQLite database to store upgrade paths, costs, and tier requirements across all player archetypes. The schema captures:

• Upgrade costs by archetype and tier level

• Attribute thresholds required for tier progression

• Attribute increase requirements for each upgrade path

  
  

Scoring Model Development

The breakthrough was building a scoring system that weighs cost vs. benefit while accounting for diminishing returns:

Efficiency Score = 100 - (SP Weight × SP Cost + Attribute Weight × (Attribute Gain × Difficulty Modifier))

Key components:

• SP Weight/Attribute Weight: User-configurable to prioritize cost vs. performance

• Difficulty Modifier: 1 + (Final Attribute / 100) to account for diminishing returns at higher tiers

• Normalization: Scores scaled 0-100 for intuitive interpretation

  
  

Difficulty Adjustment Problem

The major challenge was that not all attribute increases are equal. A +5 increase from 60→65 is fundamentally different from 90→95. My solution applies a difficulty multiplier based on the final attribute value, ensuring high-tier upgrades are appropriately penalized for their increasing marginal costs while preserving their true strategic value.

Weighting System Design

Rather than forcing a single optimization approach, I implemented user-configurable weights:

0.0 (Attribute-focused): Prioritizes attribute increases regardless of cost

0.5 (Balanced): Optimal for most players seeking value upgrades

1.0 (SP-focused): Maximizes progression per SP spent

Data Integrity Considerations

Analysis revealed scores naturally clustering in the 71-100 range. Rather than artificially expanding this to fill 0-100, I preserved the natural distribution because it authentically represents real upgrade efficiency patterns and maintains user trust through transparent, meaningful scoring.

This model revealed that mid-tier upgrades (Silver → Gold) often delivered better value than premium upgrades, which was essential for users determining if upgrading to higher tiers was worth it.

—————————

The Solutions & Product Features

AI Assistant Integration

Players needed instant answers to complex upgrade questions like "What's the best way to spend 15 SP on a Speedster?" without manually comparing hundreds of upgrade combinations.

I built an AI assistant that connects directly to the efficiency scoring model, turning natural language questions into data-driven recommendations. When users ask complex queries about budget constraints, archetype synergies, and tier planning, the AI processes these against the same algorithms powering the visualizations.

Technical Implementation:

• Integrated OpenAI API with the efficiency scoring database

• Enabled multi-factor analysis through conversational queries

• Maintained consistency with the visual tool's recommendations

Players can now get expert-level upgrade advice through simple questions, making advanced optimization strategies accessible to casual users who don't want to analyze charts.

Upgrade Efficiency Analyzer

I built a dynamic scoring system that weighs cost vs. benefit while accounting for diminishing returns at higher tiers. Users can adjust the weighting to match their strategy - some want maximum bang for their buck, others want to see attribute increases regardless of cost.

Technical Implementation:

• Difficulty adjustment algorithm that penalizes high-tier upgrades appropriately

• User-configurable weighting system (0.0 = stat-focused, 1.0 = budget-focused)

• Three visualization modes revealing different strategic insights

• Interactive charts that surface counterintuitive optimization patterns

Efficiency vs SP Cost

Efficiency vs Attribute Increase

SP Cost vs Attribute Increase

Tier Progress Visualization

The upgrade planner shows visual progress bars for each tier, with SP costs and attribute thresholds clearly marked. Users can simulate upgrade sequences before committing resources.

—————————

Design Decisions

Looking at the interface, I designed it to feel like a premium gaming analytics tool rather than a generic business dashboard:

• Signature neon green branding that immediately signals "lab" and "analysis" while maintaining gaming credibility

• Dark theme with strategic accent colors: The interface uses deep blacks with electric green highlights, creating visual hierarchy without overwhelming the data

• Ability grid visualization: Color-coded squares (bronze, silver, gold, platinum) provide instant visual feedback on upgrade status and tier progression

• SP cost displayed large and clear, making budget management intuitive

• Three-column layout: Current tier, target tier, and SP summary create a logical left-to-right progression story

• Color-coded tier dropdowns: Each tier level has its own color (bronze, silver, gold, platinum) for instant recognition

Interactive Design Elements

• Real-time SP calculations: As users select upgrades, costs update immediately in the right panel

• Hover feedback: The ability grid provides immediate visual response to user interaction

• Weighted slider control: The prominent green slider lets users adjust efficiency calculations dynamically

The result feels like a professional esports analytics tool that serious players would actually want to use, bridging the gap between data analysis and engaging user experience.

—————————

Impact & Insights

Key Findings That Changed How Players Approach Upgrades

• Archetype Efficiency: Speedsters deliver 23% higher attribute gain per SP compared to Power archetypes

• Tier Optimization: Silver → Gold transitions offer 34% better efficiency than Gold → Platinum on average

• Diminishing Returns: Upgrades beyond 90 attribute points require exponentially more SP for marginal gains

User Behavior Changes

The tool shifted user strategy from intuition-based to data-driven decision making. Dynasty players began using efficiency metrics to better determine how to upgrade their team to give themselves a competitive advantage.

—————————

Technical Stack & Implementation

• Frontend: VSCode, Python, Streamlit for rapid prototyping and deployment

• Data Processing: Pandas for ETL and analysis

• Database: SQLite for lightweight, portable data storage

• Visualization: Plotly for interactive, publication-quality charts

• AI Integration: OpenAI API for contextual recommendations

• Deployment: Streamlit Cloud for accessible demo hosting

  
  

—————————

Outcomes & Reflection

This project showcases how I can:

• Identify product opportunities in existing user workflows

• Build scalable data models that balance complexity with usability

• Design intuitive visualizations that drive decision-making

• Integrate AI capabilities to enhance rather than replace human judgment

The combination of quantitative analysis, visual design, and product thinking mirrors the skills needed for modern BI and product roles. I approach every project by understanding user needs first, then building the minimum viable solution that delivers maximum insight.

—————————

Project Links

• GitHub Repository: https://github.com/jemarisapp

• Live Demo: https://dynastylab.streamlit.app