Behind every city budget, every utility bill, or every family services case file, you'll find the same thing: everyday New Yorkers trying to navigate a complicated city. But as these massive agencies scale up, things fall through the cracks. The paperwork looks fine in an office downtown, but on the street, people deal with the friction.
I'm Jonathan Martinez. I'm a data analyst who focuses on how our city actually works. My background combines design problem-solving frameworks with deep-dive technical skills in Python, SQL, and AWS cloud architecture. I don't just stare at clean spreadsheets — I look for the anomalies, track down hidden operational costs, and build the clear, interactive dashboards teams need to stop bleeding resources.
As public and civic systems grow more complex, the gap between system logic and human interpretation widens. JM The Creative focuses on bridging this gap through structured analysis and persistent orientation, so that human intent stays the governing principle.
Heavy users collect more anomalies, not always worse systems.
The hidden cost of systems becoming harder to trust or explain over time.
The point where broken routine or repeated friction makes return unlikely.
Trust built through repeated reliability and depleted through unexplained strain.
When explainable patterns still feel costly, unfair, confusing, or suspicious.
Looking for the records, behaviors, or moments where explanation comes first.
Business-analysis studies on how public systems become costly, trusted, misunderstood, or hard to interpret — published as interactive Hex dashboards, the same evidence objects used in stakeholder review.
The specific artifacts produced during analysis to make complex system issues legible — to clients, stakeholders, and the public.
Question, method, evidence, interpretation, limits, and reproducibility.
Prioritization layers for where explanation or deeper review should happen first.
Compact views of exception shares, cost spread, and frequency variation.
Visual interfaces for inspecting friction, trust risk, and interpretation gaps.
System maps that show relationships, layers, and diagnostic logic.
Partial technical views that prove structure safely without exposing PII.
Alongside my analysis practice, I develop and maintain Orient UI — a strict conversational interaction layer for AI assistants. It acts as a visual systems analyst and human-in-the-loop facilitator: cognitive legibility, visible state, and structured surface rendering. Engineered around one rule: persistent orientation, progressive expansion.
Learn more about Orient UI →The canonical boot emblem rendered when Orient UI first activates in a terminal- or HUD-style context. Tone: mythic-technical.
Boot sigil · Activation sequence · Surface activation mark
Minimal · Structured · Expanded · Environmental
Focus · Mode · Clarity · Open Loops · One-Question Rule
User owns meaning · Assistant may suggest · State honesty
Nine cognitive surfaces · smallest surface that fits the load
Escalation & decay · Anti-patterns · Consistency rules
Interviews, clarification, one-question pacing
Comparisons, prioritization, scoring, tradeoffs
Flows, relationships, sequences, systems
Screens, layouts, dashboards, concepts
Combined views when one surface won’t do
User profiles, stakeholder framing
Says / Thinks / Feels / Does
Clarity · Compassion · Impact · Feasibility
Synthesis, consulting-style recommendations
For direct integration into Claude and other AI assistant runtimes — forces conversational models to abide by strict orientation, HITL checkpoints, and structured surface rendering.
For environmental and system configuration — clean, standardized interaction boundaries. Drop into any AI pipeline as a portable config.
Token-Optimized Object Notation — a lightweight data structure engineered to squeeze maximum context into minimal token spaces without UI bloat.
The user owns meaning, priority, and final decisions; the assistant may reflect, suggest structure, and surface tradeoffs — but never finalizes or stores anything without explicit confirmation.
Rendering scales from a single sentence to full diagnostic panels — minimal, structured, expanded, or environmental — so the visual layer adds orientation without ever becoming noise.
A concise translation of the practice into role fit, tools, strengths, and proof.
JM The Creative is the business analysis practice of Jonathan Martinez. It represents a pivot from traditional product design to business analysis on public and civic data, applying design's problem-solving frameworks to the systems that shape public life.
The practice is also the development home of Orient UI (formerly Context OS), a side framework for building cognitively legible AI interactions that preserve human agency and orientation.