publicdata.au publicdata.au on GitHub

Explore Australian Road Deaths Database, fatalities

Charts and tables over the 59,169 rows of the latest version, worked out in your browser from the Parquet file. Nothing you do here leaves the browser until you save a link.

The explorer needs JavaScript and a browser from the last few years. The dataset page has every file and the query API.

Data: Bureau of Infrastructure and Transport Research Economics, Australian Road Deaths Database, https://catalogue.data.infrastructure.gov.au/dataset/australian-road-deaths-database, accessed 26 October 2026, licensed under CC BY 4.0. Licence terms.

How it works

Change a panel

Press Edit at the top right of a panel to open its settings. Drag fields into Group by, Split by, Sort by and Filter, pick a chart type at the top, and choose sum, count or average beside each number. More fields lists everything else in the dataset. The chart types include bars, lines, areas, scatter plots, heatmaps, treemaps and sunbursts. Add a chart puts a new panel on the dashboard. Drag a panel's title bar to move it, or drag the line between panels to resize.

Filter every panel at once

The chart marked Click to filter the others does just that: click a bar and the other panels show only those rows. Click it again to clear. Any panel can do the same. Open the menu at the left of its title bar and turn on filtering from there.

Share or keep a dashboard

The address bar holds the whole dashboard, so copying it is enough. Save a short link to store the dashboard on this site and get a shorter address. A saved dashboard never changes; saving again after an edit makes a new link. Embed gives a frame you can paste into a web page, with the attribution inside it.

Download what you see

The export button in a panel's settings toolbar downloads the panel as CSV, JSON or Arrow, or a chart as PNG. The copy button beside it puts the same data on the clipboard. The full files are on the dataset page.

What the fields mean

The explorer names each field in words. The files and the query API use the field name beside it.

In the explorerField nameTypeNote
Crash IDcrash_idstringThe publisher's crash identifier. Every person killed in one crash shares it.
StatestatestringNSW, VIC, QLD, SA, WA, TAS, NT or ACT.
Monthmonthinteger
Yearyearinteger
Day of weekday_of_weekstring
TimetimestringTime of the crash as HH:MM:SS.
Crash typecrash_typestringSingle or Multiple vehicle, or Unknown.
Bus involvedbus_involvementboolean
Heavy rigid truck involvedheavy_rigid_truck_involvementboolean
Articulated truck involvedarticulated_truck_involvementboolean
Speed limit (km/h)speed_limitintegerPosted speed limit in km/h. The publisher's -9 for unknown is null here.
Road userroad_userstringDriver, Passenger, Pedestrian, Motorcycle rider, Motorcycle pillion passenger, Pedal cyclist or Unknown.
Gendergenderstring
AgeageintegerAge in years. The publisher's -9 for unknown is null here.
Remotenessnational_remoteness_areas_2021string
ABS area (SA4)sa4_name_2021string
Council areanational_lga_name_2021stringCouncil area on the 2021 boundaries.
Road typenational_road_typestring
Christmas periodchristmas_periodboolean
Easter periodeaster_periodboolean
How it runs

The page downloads DuckDB and Perspective once, about 17 MB, then this version's data.parquet (874 KB). Every chart is a query DuckDB answers in the browser, so there is no rate limit and the files stay the only source. A saved dashboard over a dated version keeps showing that version's numbers.