This project analyzes road safety data from Victoria, Australia, focusing on the relationship between temporal patterns, speed zones, and geographic risk factors. It transforms raw crash data into actionable insights for infrastructure policy and public safety.
- Friday Peak: The 14:00–18:00 window consistently records the highest frequency of accidents, peaking on Friday afternoons with 790 incidents.
- Commute Patterns: Weekdays exhibit a bimodal pattern (morning/afternoon peaks), while weekends show higher crash proportions during night periods.
- Environmental Risk: Dusk and Dawn periods (06:00–07:00 and 17:00–18:00) present high transitional risks due to rapid visibility changes and glare.
-
Severity Scaling: Injury severity rises almost monotonically with speed, reflecting the physical law
$E_k = \frac{1}{2}mv^2$ , where doubling speed quadruples impact energy. - Volume vs. Risk: The 60 km/h zone has the highest absolute victim count (62,887) due to traffic density, but the 90 km/h zone records the highest average severity at 1.60 victims per crash.
- Infrastructure Mitigation: Serious injuries account for ~30% of crashes at 30 km/h, rising to ~48% at 100 km/h. A slight dip at 100 km/h suggests that freeway-grade safety barriers effectively mitigate high-speed impacts.
- Melbourne (Inner City): High accident volume but low severity (32.2%). Cyclist involvement is 3–9x higher here than in any other LGA.
- Geelong (Regional): Highest fatal/serious injury rate (50.5%) due to high-speed undivided rural roads.
- Casey & Hume (Outer Suburbs): Motorcyclists are the dominant vulnerable road user group, while cyclists are nearly absent.
- Language: Python 3.8+
- Libraries:
pandas,numpy,matplotlib,seaborn