Australia’s public shark-incident database holds 1,296 records, Roughly 33% of them were provoked, meaning the person in the water did something to provoke the shark first. The story these records tell us is less about the shark and more the behaviour of the people who were subjected to the shark attack.

1. Something Changed in the Water

Recorded shark incidents have surged, but fatal ones have not

Figure 1

Recorded Shark Incidents per year in Australia, 1990 -2025

Note: Counts include every incident in the database with a recorded year; 2026 is excluded becuase the year is incomplete.since 2000, about 26% of recorded incidents involved no injury. Data from the Australian shark-incidents Database (public version).

2. What are they doing in the Water?

Across the years the surge in activities have changed with boarding taking a sharp increase from the 2000s.

Figure 2

Average Shark Incidents per Year by Victim Activity and Decade, 1900-2025

Note. Each point is the average number of incidents per year in that deacde. The 2020s cover2020-2025 only. Activities are grouped from the database categories(boarding included surfing).Incidents with other or unknown activities (n =107) are not shown. Use the menu above the figure to highlight one activity. Data from the Australian Shark Incident Database (Public version).

About third of the total incidents and every spearfishing incident was provoked.

Figure 3

Share of Shark Incidents That Were Provoked by Victim Activity

Note. Includes all records from 1791 to 2026 with a know provocation status (6 records excluded). Provoked and unprovoked follow the database's own classification. The dashed line indicates the share across all activities (33%). Hover over a bar for counts and fatal incidents, and use the menu to change the sort order. All colors are consistent with figure 2. Data from the Australian Shark-Incidents database (Public version).

4. Different coasts, diffferent sharks

Which shark is involved depends on the coast: White sharks mostly in the south, tiger and bull sharks further north.

Figure 4

Shark Incidents by Species Across Australian Coastline

Note. Each point is one incident (n =1,295) at its recorded location, nudged by up to about 2km so overlapping points stay visible.An orange-red halo marks a fatal incident.The map is cumulative. The slider shows all records up to the end of the selected decade, and the 2020s frame includes every mapped record to 2026. Records without usable coordinates (n = 1) are not mapped. Other and unknown species (18% of mapped incidents) includes species outside the four named groups and records with no species recorded. Press play or drag the slider, and use the buttons to zoom to a state or territory. Data from the Australian Shark Incidents Database (public version)Basemap: Esri World Ligh Gray Canvas.

5. Few incidents are deadly, but is it safer per person?

Fewer incidents are fatal, but the fatality rate per person (risk of dying) has not Fallen

Figure 5

Share of unprovoked Shark Incidents That were Fatal and Fatal incidents per Million People, by Decade

Note. Based on 555 unprovoked incidents with a known outcome, from 1960 to 2025 (76 fatal). Provoked incidents are excuded beacuse activities such as spearfishing are almost all provoked and rarely fatal, which would distort the comparison. Points show each decade; vertical bars show 95% confidence intervals (exact binomial in Panel A, exact Poisson in Panel B), and wide bars mean few fatal incidents. The 2020s cover 2020-2025 only.Rates use Australia's estimated residents population, interpolated between 30th June estimates for 1960, 1970, 1980, 1990, 2010, 2020, 2025. Click a decade in either panal to highlight it in both, and double click to clear. Data from the Australian Shark Incidents Database (public version) and the Australian Bureau of Statistics

What does this mean ?

Across six decades of Australian shark incident records, one main insight stands out. The fatality of the shark attack has improved but the likely hood of a fatal attack occurring has not. The share of unprovoked incidents that were fatal fell from about one in four in the 1960s to 1980s to about one in ten in the 2020s and 2020s, yet fatal incidents per million have not fallen with it. The 2020s show the highest rate of any decade since 1960. The records also show that a third of the incidents were provoked and that the activities people were doing have changed. Over the decades, only a few fatal incidents have be recorded, but that does not mean the water is safe.

Limitations

Early records are likely to be skewed towards fatal or dramatic events. Minor incidents are far more likely to be reported today. Victim age is only about 60% complete, therefore it is not used. “Provoked” is the database’s own classification

Acknowledgement/ Respect for people

Fatal incidents involve real people. The reports says “incident” rather than “attack” and does not name any individual or have any graphic images.

Gen AI acknowledgement

Tool used – Claude (Anthropic), Accessed via application.

Claude was used for simple debugging purposes of the R code and knitting the document to HTML. All code fixes produced by Claude evaluated based on the error message output from R and verified based on my knowledge of R and the output produced. I confirm that I have revised and tested and take full responsibility of the code edits done by Claude. All brainstorming and constructing the visualisation are my own based on the knowledge gathered from the course materials. (References attached below).

References

[1]. Australian Shark Incident Database | Taronga | Taronga Zoo. (n.d.). https://www.taronga.org.au/wildlife-and-conservation/the-institute/australian-shark-incident-database.

[2]. Australian Bureau of Statistics. (2012). Year book Australia, 2012 (Cat. no. 1301.0). https://www.abs.gov.au/AUSSTATS/abs@.nsf/7d12b0f6763c78caca257061001cc588/f91636667dabbfbeca257a07001ac43e!OpenDocument.

[3]. Australian Bureau of Statistics. (2020, December 17). Australia’s population grew by 1.3 per cent [Media release]. https://www.abs.gov.au/media-centre/media-releases/australias-population-grew-13-cent.

[4]. Australian Bureau of Statistics. (2025, December 18). Population up 1.5% in June 2025 [Media release]. https://www.abs.gov.au/media-centre/media-releases/population-15-june-2025.

[5]. Baglin, J. (2026, March 26). Chapter 1 Design and Integrity | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/design-and-integrity

[6]. Baglin, J. (2026b, March 26). Chapter 3 Visual Perception and Colour | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/visual-perception-and-colour.html.

[7]. Baglin, J. (2026c, March 26). Chapter 4 Accessibility and Deception | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/accessibility-and-deception.html.

[8]. Baglin, J. (2026d, March 26). Chapter 5 Grammar and Vocabulary | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/grammar-and-vocabulary.

[9]. Baglin, J. (2026e, March 26). Chapter 6 Multivariate Strategies | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/multivariate-strategies.html.

[10]. Baglin, J. (2026e, March 26). Chapter 7 Spatial Data | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/spatial-data.html.

[11]. Baglin, J. (2026e, March 26). Chapter 8 Adding Interactivity | Data Visualisation: From Theory to Practice. https://data-visualisation.stem.melbourne/adding-interactivity.html.

[12]. Anthropic(2021). Rmd knit to HTML, Claude, https://claude.ai/share/2a4448fc-0084-414e-8182-8743473b6d32