Refining a visualisation prompt

Name:

Date:

Class:

Quick Questions

Shade in the A or B circle for each question.

1. An AI draws a chart with no axis labels. Which follow-up prompt refines it best?

A Add the x-axis label 'Month' and the y-axis label 'Sales ($)' to the chart.

B Make the chart better.

2. The AI made a pie chart, but you need to compare 8 regions. Which follow-up is better?

A I don't like this chart. Try again.

B Change the pie chart to a bar chart so the 8 regions can be compared by length.

3. True or false: after you send one follow-up prompt, the visualisation is finished and does not need to be checked again.

A False: review the new chart and repeat the cycle until it is right.

B True: one follow-up always fixes everything.

4. Which follow-up adds detail to a line chart of rainfall?

A Add data labels showing the rainfall in mm above each point on the line.

B Add more stuff to the chart.

5. What should you do first when refining an AI-generated chart?

A Write a brand new prompt without looking at the chart.

B Review the chart and note exactly what is wrong or missing.

Match the Pairs

Draw a line connecting each item on the left to its match on the right.

Left column		Right column
1Change the chart to a line chart.		AA follow-up prompt that changes the chart type to suit the data
2Label the y-axis 'Rainfall (mm)' and add a chart title.		BA follow-up prompt that fixes labels and the title
3Refining a prompt by adding detail after reviewing the chart		CPrompt 1 > check chart > Prompt 2 > check chart > Prompt 3
4Iterating: send a follow-up, check the new chart, then repeat until it is right		DToo vague. It does not name what to change, so the AI has to guess.
5Make it better.		EThis skips reviewing the chart and throws away what already works instead of naming what to change.
6Ignore the chart and write a whole new prompt from scratch every time.		FAdd data labels to each column and show only Year 12 results.

Definitions

Refining a prompt

Fill in the Blanks

Complete each line of code by filling in the missing parts.

Compare 8 classes: the pie is too crowded. Change the pie chart to a ___ chart.

Problem: months are shown, but no label. Follow-up: add an ______ label 'Month'.

Bars show sales, but no values are visible. Add data ______ to the top of each bar.

Review the chart, send a follow-up, then ______.

Show how study hours relate to test marks. Change the line chart to a _______ chart.

Spot the Error

Shade in the letter of the line that contains an error.

Question 1

A review: y-axis has no label

B fix = "Label x-axis as Day"

C chart = ai(fix)

Question 2

A chart = ai("Pie chart: sales")

B review: too many slices

C fix = "Add more slices"

Question 3

A chart = ai("Line chart: temp")

B review: chart has no title

C fix = "Add a legend"

Question 4

A chart = ai("Bar chart: sales")

B review: axis ends 50, max 80

C fix = "Use a smaller font"

Question 5

A chart = ai(first_prompt)

B review(chart)

C while not chart_ok:

D fix = "Try again"

E chart = ai(fix)

F review(chart)

Answer Key

Refining a visualisation prompt

Section A: Quick Questions

A — Add the x-axis label 'Month' and the y-axis label 'Sales ($)' to the chart.

B — Change the pie chart to a bar chart so the 8 regions can be compared by length.

A — False: review the new chart and repeat the cycle until it is right.

A — Add data labels showing the rainfall in mm above each point on the line.

B — Review the chart and note exactly what is wrong or missing.

Section B: Match the Pairs

Change the chart to a line chart. → A follow-up prompt that changes the chart type to suit the data

Label the y-axis 'Rainfall (mm)' and add a chart title. → A follow-up prompt that fixes labels and the title

Refining a prompt by adding detail after reviewing the chart → Add data labels to each column and show only Year 12 results.

Iterating: send a follow-up, check the new chart, then repeat until it is right → Prompt 1 > check chart > Prompt 2 > check chart > Prompt 3

Make it better. → Too vague. It does not name what to change, so the AI has to guess.

Ignore the chart and write a whole new prompt from scratch every time. → This skips reviewing the chart and throws away what already works instead of naming what to change.

Section C: Definitions

Refining a prompt — Reviewing the AI's chart, then sending a follow-up prompt that names exactly what to change, such as labels, chart type or detail, and repeating until it is right.

Section D: Fill in the Blanks

Compare 8 classes: the pie is too crowded.
Change the pie chart to a bar chart.

Problem: months are shown, but no label.
Follow-up: add an x-axis label 'Month'.

Bars show sales, but no values are visible.
Add data labels to the top of each bar.

Review the chart, send a follow-up, then 
repeat.

Show how study hours relate to test marks.
Change the line chart to a scatter chart.

Section E: Spot the Error

A — The follow-up asks to label the x-axis, so the review must name the x-axis as the problem. Reviewing the chart first tells you exactly what to name in the follow-up. It should read: review: x-axis has no label.

C — The review found too many slices, so adding more makes the chart worse. A follow-up that changes the chart type should name the new type. It should read: fix = "Change to bar chart".

C — The problem found in the review was a missing title, but this follow-up asks for a legend. A follow-up must name the exact fix for the problem you found. It should read: fix = "Add title: Daily Max".

C — The review found the y-axis cuts off the data, so changing the font does not fix it. The follow-up must name the axis change. It should read: fix = "Extend y-axis to 100".

D — Iterating means each follow-up names exactly what to change. 'Try again' is vague, so the AI cannot know what to fix. It should read: fix = name_change().