Figures and maps - SDS320 Spatial Data Analytics

Communicating spatial results clearly and honestly

1. Why figures and maps matter¶

Figures and maps are evidence for your project argument.

A good figure helps the reader understand what you analysed, what you found and why it matters. A weak figure can confuse the reader, hide uncertainty or make a result look more precise than it is.

Start every visual with a question:

What should the reader learn from this figure?
Where is the result located?
Why does this visual matter?
Who is the audience?

2. What makes a good project figure?¶

A good figure or map usually:

supports the research question,
focuses on one main message,
has a clear caption,
shows enough spatial context,
includes legend, labels, units and time frame where needed,
uses suitable colour choices,
is readable in report and presentation formats,
avoids unnecessary clutter,
makes limitations visible or discusses them in the caption or text.

3. Types of useful project visuals¶

Different project stages need different visuals.

Visual type	Use it for
Overview map	Show the study area and spatial context.
Input data map	Show what the source data look like.
Preprocessing or quality-check figure	Show clipping, masking, cloud problems, missing data or alignment checks.
Workflow diagram	Explain the steps from raw data to result.
Model output map	Show predictions, detections, segmentation masks or regression outputs.
Comparison map	Compare dates, methods, classes, scenarios or before/after outputs.
Error or uncertainty figure	Show where predictions are uncertain or where errors occur.
Summary chart	Show counts, areas, class proportions, metric values or trends.
Key-result figure	Communicate the main answer to the research question.

Not every project needs all of these. Choose figures that serve your argument.

4. Map design checklist¶

Before using a map in your report or presentation, check:

Does the map answer or support a specific question?
Is the study area clear?
Is the legend readable?
Are labels, units and time frames included where needed?
Is the projection suitable for the purpose?
Is the colour scale appropriate for the data type?
Are class breaks consistent when comparing maps?
Are uncertainty or limitations mentioned where relevant?

For satellite RGB or false-colour images, explain the assigned channels, for example:

R, G, B = shortwave-infrared, near-infrared, red

5. Captions and interpretation¶

A figure should not stand alone without explanation.

A useful caption usually states:

what is shown,
where and when it is shown,
which data or method produced it,
what the reader should notice,
any key limitation needed for interpretation.

For report writing, figure captions are usually placed below the figure. Table captions are usually placed above the table.

A concise caption pattern is:

Figure X. [What is shown] for [area/time period] based on [data/method]. [Main pattern or interpretation]. [Limitation if needed].

Example:

Figure X. Predicted built-up areas in the study area based on a semantic segmentation workflow applied to Sentinel-2 imagery. The largest predicted built-up clusters occur near the main transport corridor. Small isolated predictions should be interpreted carefully because no independent validation labels were available.

6. Visual hierarchy and storytelling¶

Visual hierarchy means that the most important information is easiest to see. Use hierarchy through title, caption, colour emphasis, annotation, spacing and figure order.

Storytelling does not mean exaggeration. It means guiding the reader from question to evidence to interpretation.

For your final presentation, one strong key-result figure is often more effective than many weak figures.

7. Potential pitfalls¶

Pitfall	Better approach
The figure does not support the research question	Remove it or change the message.
The map has no spatial context	Add study area outline, reference locations or overview map.
Labels or legends are hard to read	Simplify classes and increase readability.
Colours imply false precision	Use a colour scale that matches the data type.
Maps are compared with different class breaks	Use consistent breaks where comparison is intended.
Model output is shown without limitations	Add evaluation, uncertainty or a clear caution.
Caption only describes colours	Explain what the figure demonstrates.

8. Mini task¶

Choose one figure or map that could become central to your project.

Figure purpose:
Research question connection:
Data shown:
Method or processing step:
Main message:
Important limitation:

Then draft a two-sentence caption.

9. Key takeaways¶

Figures and maps should support your project argument.
Start every visual with a question and audience.
Keep maps readable, honest and focused.
Captions should explain what the figure demonstrates and what it does not show.
A few strong visuals are better than many unfocused ones.