© 2026, Ian Muehlenhaus
Prompt Cartography

Labs / Chapter 4

Lab 4: Build a Prompt-to-Map Pipeline
Sequence data, visual, interface, specification, critique, and publishing decisions

Beginner ~90–120 min Chapter 4 Core PDF

Chapter 4 turns prompt cartography into a workflow. In this lab, you will build a pipeline packet that names when key decisions happen, what each stage needs, and how later critique can trace output back to earlier intent.

Purpose

~5 min

Chapter 4 argues that pipelines do not replace cartographic judgment. They make judgment visible, sequenced, and easier to critique. In this lab, you will turn a map idea into a prompt-to-map pipeline packet that can guide later LLM-assisted work without rushing into execution.

Core ideaA pipeline is not a factory. It is a named sequence of design decisions that keeps speed from collapsing meaning into execution.

Learning Objectives

~3 min
  • Explain the difference between a workflow and a pipeline.
  • Create a data brief that anchors purpose, audience, scope, limitations, and ethical concerns.
  • Separate visual planning from styling and interface planning from feature collection.
  • Write a specification as a flexible contract rather than a command list.
  • Use critique prompts to reconnect output to earlier intent.
  • Explain publishing as accountability rather than merely deployment.
ToolsA web browser, one free LLM, and a notes document. No coding or GIS software required.

Setup: Choose a Map Idea Worth Sequencing

~7 min

Choose a map idea where purpose, data limits, interface choices, and accountability matter. Keep this topic for the full lab.

  • Public cooling centers during heat waves
  • Traffic safety near schools
  • Emergency shelters and evacuation routes
  • Neighborhood access to clinics, parks, or groceries
  • Environmental risk communication for a local audience
Setup deliverableWrite one sentence naming your topic, intended audience, and what the map should help that audience decide, understand, or do.

Part 1 — Draft the Data Brief

~18 min

The data brief asks what the map can responsibly claim before anyone starts designing.

Data Brief
Map topic:
Intended purpose:
Audience:
Spatial scope:
Temporal scope:
Key variables needed:
Claims the map can support:
Claims the map cannot support:
Known limitations or uncertainty:
Ethical considerations:
One clarifying question before data work begins:

After you draft it, ask an LLM to critique the brief.

Act as a skeptical cartographic project reviewer. Critique this data brief before any map is designed. Identify vague claims, missing scope limits, ethical risks, and one question I should answer before moving to visual planning.
Mini-deliverableYour revised data brief and one sentence naming the most important limitation.

Part 2 — Visual Planning Before Styling

~15 min

Visual planning is about hierarchy, emphasis, and tone before colors, fonts, or symbols.

Using my data brief, help me create a visual planning note. Do not choose final colors or symbols yet. Instead, specify:
1. what should feel visually central
2. what should recede
3. what tone the map should carry
4. what visual defaults would be misleading
5. what the user should notice first, second, and last

Revise the response so it is written in your voice, not the model’s.

Part 3 — Interface Planning: Say No on Purpose

~15 min

Interface planning anticipates use. It is not a shopping list of widgets.

Using my data brief and visual planning note, create an interface planning note for this map. Include:
- primary user task
- likely device or context
- expected time and attention
- whether exploration should be guided or open-ended
- 2 interactions that support the purpose
- 2 interactions I should reject, with reasons
ReminderInterface choices are rhetorical choices. A filter, slider, popup, or panel changes what users think the map is asking them to do.

Part 4 — Write the Specification as a Contract

~20 min

Now combine the earlier stages into a specification. The goal is to stabilize intent without micromanaging the LLM.

Prompt-to-Map Specification Contract

Data brief summary:
[Paste concise version]

Visual planning commitments:
[Paste concise version]

Interface planning commitments:
[Paste concise version]

System freedom:
The LLM may suggest implementation approaches, but must preserve the stated purpose, audience, scope limits, visual hierarchy, and interface rationale.

Non-negotiables:
- [Constraint 1]
- [Constraint 2]
- [Constraint 3]

Flexible choices:
- [Choice 1]
- [Choice 2]

Output requested:
Return a map concept brief and implementation plan. Do not write code.

Run the contract through an LLM and save the response as your pipeline output.

Part 5 — Critique the Pipeline Output

~20 min

Critique is a design stage, not a cleanup phase. Ask the LLM to reconnect the output to the earlier pipeline stages.

Critique the map concept and implementation plan against my pipeline packet. Identify:
1. where it aligns with the data brief
2. where visual planning was respected or ignored
3. where interface suggestions support or distract from intended use
4. where the specification was too vague
5. where the output drifted from my intent
6. one revision I should make before publishing or building
Mini-deliverableRevise one earlier pipeline stage based on the critique. Do not only revise the final output.

Part 6 — Publishing Accountability Note

~10 min

Publishing is where delegation ends and responsibility crystallizes. Write a short accountability note before calling the concept ready.

Publishing Accountability Note
If this map shipped with my name on it, I would be comfortable with:

I would hesitate because:

The single revision that would most improve accountability is:

The audience most likely to misread this map is:

The claim this map must not be allowed to imply is:

Submission

~15 min

Submit or save the following:

  • Revised data brief
  • Visual planning note
  • Interface planning note with rejected interactions
  • Specification contract
  • Pipeline output and critique
  • Publishing accountability note
  • A 300-500 word reflection on how sequencing changed the quality or traceability of your decisions
How to Do Well A friendly self-check: aim to show clear sequencing, specific data and scope limits, visual and interface decisions tied to purpose, a contract-style spec, and critique that revises the pipeline rather than only the final answer. Think of this as the success path while you work, not a gotcha at the end.

Psst! Before You Turn This In...

~3 min

Can you tell Ian thinks this is really, really important yet? Good. Keep going.

Do the tiny-but-mighty judgment check. If these answers are fuzzy, revise the map idea before polishing the prompt.

  • Purpose: Can you say what this map helps someone understand or decide?
  • Evidence: Can you point to the data, source, or context behind the claim?
  • Omission: Did you leave out, downplay, or defer anything that would distract or mislead?
  • Risk: What could be overstated, exposed, or misunderstood?
  • Human check: What must be verified outside the LLM?
  • Stop rule: What would make you redesign, withhold, or simplify this map?
Ian's recurring refrain Prompt frameworks are scaffolds. The LLM can suggest, critique, and surprise you; you still own the cartographic judgment.