A0: Applied Exercise

Artefacts Checklist

Use this checklist to keep track of the working artefacts and supporting material you should produce throughout the Applied Exercise.

The aim is to make sure your investigation remains complete, traceable, and easy to review without requiring you to document every minor step. Each section below identifies the main material to retain or submit for GATHER → EXAMINE → IDENTIFY → INTERROGATE → EVALUATE → TRANSFORM → SELECT.

Where specific filenames, quantities, or folder structures are required, they are shown directly in the relevant section.

1. GATHER

For each of your approximately 5 site collections, submit:

• Approximately 1–2 Context views / representations.
• Approximately 3–5 System views / representations.
• Approximately 4–5 Detail views / representations.
• 1 nested visual explanation showing how selected Details relate to the System, and how the System relates to its wider Context.
• The original photographs / visual material used to form the collection.

Visual Collection

Across approximately 5 collections, you should therefore have:

In total, your GATHER material should contain approximately 45–60 views / representations, together with 5 nested visual explanations.

For photographs:

• System photographs: provide the corresponding EXIF CSV file for each photograph.
• Detail photographs: EXIF CSV files are optional.
• Context photographs: EXIF CSV files are not required.

For plans, scans, drawings, screenshots, renderings, or other non-camera material, EXIF extraction is not required.

Always retain and submit the original image or source file. Do not resize, compress, crop, convert, screenshot, or otherwise re-save original photographs before preserving them.

For detailed instructions on taking, exporting, and preserving photographs, see [Taking & Exporting Photos].

Nested Visual Explanations

Combine your 5 nested visual explanations into a single PDF:

UID_A0_Gather_Nested.pdf

Use approximately one page per collection.

Each page should clearly show the relationship:

Context → System → Detail

Use simple crops, boxes, arrows, labels, or annotations so that selected Details can be located within the System, and the System within its wider Context.

File Structure

Organise your GATHER material as follows:

01_GATHER/
│
├── C01/
│   │
│   ├── C01_Context_01.HEIC
│   ├── C01_Context_02.HEIC
│   │
│   ├── C01_System_01.HEIC
│   ├── C01_System_01_EXIF.csv
│   ├── C01_System_02.HEIC
│   ├── C01_System_02_EXIF.csv
│   ├── ...
│   │
│   ├── C01_Detail_01.HEIC
│   ├── C01_Detail_02.HEIC
│   └── ...
│
├── C02/
│   └── ...
│
├── C03/
│   └── ...
│
├── C04/
│   └── ...
│
├── C05/
│   └── ...
│
└── UID_A0_Gather_Nested.pdf

If you choose to include EXIF metadata for a Detail photograph, use the same base filename:

C01_Detail_01.HEIC
C01_Detail_01_EXIF.csv

Keep the original file format produced by your device or source. .HEIC is shown only as an example.

2. EXAMINE

Submit one DOCX or PDF containing your observations.

General Observations

Briefly note:

• Important or recurring features.
• Common uncertainties, ambiguities, or missing information.
• Conditions that may be interesting to test with the AI.

Individual Site Collections

For each site collection (C01–C05), briefly record:

• What you think you are seeing.
• Important features or conditions.
• What remains uncertain or difficult to understand.
• What may be worth testing with the AI.

You may include cropped images or simple annotations where helpful.

File Structure

Submit your file as .DOC or .PDF:

02_EXAMINE/
└── UID_A0_Examine.pdf

3. IDENTIFY

Submit one DOCX or PDF containing the questions you developed.

Systematic Questions

Develop approximately 5 questions that can be asked across multiple site collections.

Across these questions, include at least:

• 1 relationships / connectivity question.
• 1 interpretation / reasoning question.
• 1 uncertainty / missing information question.

Number them clearly:

1. Q01: …
2. Q02: …
3. Q03: …
4. Q04: …
5. Q05: …

Site Collection-Specific Questions

For each site collection (C01–C05), develop approximately 1–2 additional questions responding to something distinctive in that collection.

Across approximately 5 sites collections, this gives approximately 5–10 collection-specific questions, in addition to the 5 systematic questions.

You may include a cropped image or simple annotation where helpful.

File Structure

Submit your file as .DOC or .PDF:

03_IDENTIFY/
└── UID_A0_Identify.pdf

4. INTERROGATE

Submit one PDF documenting how the same systematic question performs across your approximately 5 site collections.

Systematic Question Across Collections

Choose one of your systematic questions and test it independently across approximately 5 collections.

For each collection, show:

• The relevant image(s) provided.
• The question asked.
• The model’s first response.

Keep the question wording as consistent as possible across the five collections.

Your PDF should therefore make the comparison easy to follow:

C01 → C02 → C03 → C04 → C05

File Structure

Submit your record as a PDF:

04_INTERROGATE/
└── UID_A0_Interrogate.pdf

5. EVALUATE

Submit one DOCX or PDF summarising how the model performed across your tests.

Systematic Questions

For each of your approximately 5 systematic questions, briefly summarise how the model performed across the collections.

You may use:

• A few words or short notes.
• Tables or simple charts.
• Check marks or other visual comparisons.
• Correct / Partially Correct / Incorrect, where useful.

Focus on the main patterns you observed across collections, rather than describing every individual response in detail.

A simple comparison table may be useful:

You may adapt the table, symbols, or categories to suit your questions. The aim is simply to make the performance of each systematic question across the site collections easy to compare.

Individual Collections

For each collection (C01–C05), briefly assess how the model performed on the questions you tested.

Note the most important:

• Successes or errors.
• Uncertainties.
• Unsupported assumptions.
• Important changes or corrections in the model’s responses.

Keep this concise. Small images, tables, or examples may be included where useful.

File Structure

Submit your evaluation as DOCX or PDF:

05_EVALUATE/
└── UID_A0_Evaluate.pdf

6. TRANSFORM

Apply one shared transformation strategy across at least 3 selected site collections.

Transformation Criteria

Submit a short DOCX or PDF stating, in a few words:

• The shared transformation intention / strategy.
• What you are trying to change, improve, explore, or challenge.
• Any important constraints or qualities to preserve.

This should define one consistent transformation approach that can be applied across all 3 selected collections.

Transformed Outputs

For each selected collection:

• Retain the starting artefact.
• Produce approximately 10 transformed variations using the shared transformation strategy.
• Save the transformed outputs individually.

Across at least 3 collections, you should therefore produce approximately 30 transformed outputs in total.

You do not need to submit every conversational exchange or intermediate prompt here.

File Structure

Submit the transformation criteria as DOCX or PDF, with transformed outputs organised by collection:

06_TRANSFORM/
│
├── UID_A0_Transform.pdf
│
├── C01/
│   ├── C01_Transform_01.png
│   ├── C01_Transform_02.png
│   ├── C01_Transform_03.png
│   ├── ...
│   └── C01_Transform_10.png
│
├── C02/
│   ├── C02_Transform_01.png
│   ├── C02_Transform_02.png
│   └── ...
│
└── C03/
    ├── C03_Transform_01.png
    ├── C03_Transform_02.png
    └── ...

7. SELECT

Submit one DOCX or PDF documenting how you evaluated and selected the final transformed outputs.

Selection Criteria

Define a small set of shared criteria that you will use across all transformed outputs.

Keep this concise.

Compare & Select

Evaluate the transformed outputs against the shared criteria.

A simple table is sufficient. Include:

• The transformed image.
• Its exact filename.
• Your selection criteria.
• Check marks, short notes, or simple scores.
• A clear indication of the approximately 3 selected outputs per collection.

For example:

Across at least 3 transformed collections, you should select approximately 9 final variations in total.

Selected File List

Also submit a plain-text file listing the exact filenames of all selected outputs, one per line.

For example:

C01/C01_Transform_02.png
C01/C01_Transform_06.png
C01/C01_Transform_09.png
C02/C02_Transform_01.png
C02/C02_Transform_05.png
C02/C02_Transform_08.png
C03/C03_Transform_03.png
C03/C03_Transform_07.png
C03/C03_Transform_10.png

Save this as a standard plain-text .txt file.

File Structure

07_SELECT/
├── UID_A0_Select.pdf
└── UID_A0_Selected_Files.txt

The transformed outputs themselves remain in 06_TRANSFORM/.

Submit the artefacts together as a single zipped file.