Reading accident data goes wrong in three predictable places, and all three show up in industry presentations: a count read as if it were a rate, causes drawn as slices of a pie, and two incompatible series on one axis. Our free accident statistics charts are built to make each of those mistakes hard to commit — which is a better argument for using them than the fact that they are free.
What the charts show
The Polish series comes from GUS, the national statistics office, via its Local Data Bank (GUS, Wypadki przy pracy). In the most recent year in the tool, 2024, it records about 67,000 people injured at work in Poland, 250 of them fatally, at a rate of 4.74 per 1,000 workers.
Four views sit on top of that series:
- The yearly trend. Injuries fell sharply from around 88,000 a year in the 2010s to the mid-60,000s after 2020, and the incidence rate per 1,000 workers roughly halved. Fatal accidents follow a slower path and move year to year with the economy and with reporting.
- By region. Absolute counts concentrate in the most industrialised and populous voivodeships, led by Silesia, Masovia and Greater Poland. The rate per 1,000 tells a different story: it is highest in industrial regions and lowest in Masovia, whose large service economy dilutes it. Reading the count when you meant the rate is the most common mistake this chart prevents.
- By industry. Manufacturing accounts for the largest single share of injured people, followed by trade and repair, health care and transport.
- By cause. In the GUS classification the dominant cause is improper behaviour of the worker, ahead of improper handling of a material agent and organisational factors.
The EU view is separate and uses Eurostat's ESAW standardised rate per 100,000 employed, which corrects for the industry mix of each economy (Eurostat, hsw_mi01).
Three caveats built into the tool
Statistics like these are easy to misuse, so the tool carries its own warnings rather than leaving them to a footnote.
Under-reporting distorts cross-country comparisons. Countries with insurance-based reporting systems record more non-fatal accidents than countries with declaration-based ones. Eurostat says so explicitly. That is why the country ranking leads with fatal accidents, which are reliably comparable, and shows non-fatal rates only with the caveat attached.
Causes are contributing factors, not shares. GUS codes each accident with more than one cause, so they sum to more than the number of accidents. A pie chart of these figures would be wrong, which is why the tool does not draw one.
GUS and Eurostat are never mixed on one chart. Polish pages use GUS figures under national definitions; cross-country charts use the Eurostat standardised rate. Putting the two series on the same axis produces a comparison that means nothing, and it happens more often than you would think in industry presentations.
What to actually do with the numbers
The honest use of national statistics is context, not proof. They tell you which hazards dominate in your sector and which direction things are moving; they do not tell you what your own accident rate is or what a specific control would change.
Two combinations work well in practice. Pair the industry chart with your own incident log to see whether your profile matches your sector or diverges from it — divergence in either direction is worth explaining. And pair a chart with a cost figure: our workplace accident cost calculator builds the cost of one event bottom-up from your records, which turns "manufacturing has the most accidents" into a number your finance team can act on.
For the training argument that usually follows, slips, trips and falls: training that actually works covers the most common category, and the VR training ROI calculator weighs a training programme against the cost of the incidents you are trying to avoid.
Embedding and citing
Every chart offers a ready-made citation and an HTML link, and the tool can be embedded in an iframe on your own site, so the chart on your page updates when the data does. Reuse is free with attribution, and the attribution matters: we present GUS and Eurostat data, we are not the primary source, and anyone checking your slide should be able to reach the original in one click.
A last word on honesty with charts: the caveats are not decoration. If a slide of ours ends up in your risk assessment, whoever reviews it should be able to see in one click which series it draws — and what that series cannot tell them.
