Queensland Crash Analysis Findings

1. Crashes by Day of Week

Hypothesis: Crashes will be most frequent on weekdays during commute hours and on Friday and Saturday nights due to social activities.
Findings: Crash counts peak on Friday at 67,446, followed by Thursday and Wednesday. Sunday recorded the lowest count at 45,785. This aligns with the typical workweek traffic flow.
Key Insight: Friday crashes are 47% higher than Sunday crashes, indicating significant weekly variation in crash risk.

Findings: Southport has the highest number of crashes, followed by Brisbane City. These are likely major commercial or transport hubs with heavy traffic flow.
Key Insight: The top 10 suburbs account for a disproportionate share of total crashes, suggesting targeted safety interventions in these areas could have significant impact.

Findings: March shows the peak crash count, likely due to increased travel during holiday season and weather conditions. January has the fewest incidents.
Key Insight: March crashes are approximately 20% higher than January crashes, indicating significant seasonal variation.

Findings: The highest crash count occurred in 2011 (23,525), while 2024 recorded the lowest (6,943). Crash numbers declined sharply after 2010.
Key Insight: The 70% reduction from 2011 to 2024 suggests significant improvements in road safety measures or traffic management.

Findings: The heatmap shows a dark red concentration during the **afternoon hours (3-6 PM) on weekdays**, with **Friday afternoon being the absolute peak**.
Key Insight: Friday 5 PM shows the highest crash concentration, suggesting end-of-week fatigue combined with rush hour traffic creates the most dangerous driving conditions.

Findings: Most high-incident suburbs follow the general trend of a Friday peak, but some show a steeper increase than others, indicating different weekly traffic rhythms.
Key Insight: Commercial districts show more pronounced weekday peaks, while residential areas show more consistent patterns.

Findings: The average number of fatalities per crash increases dramatically with speed limit. **"Darkness - Not Lighted"** is by far the most dangerous lighting condition.
Key Insight: Crashes in darkness without lighting are **3-4 times more likely** to result in fatalities compared to daylight crashes.

Findings: The interactive Folium map shows a clear clustering of crashes. **Brisbane City** recorded the highest number of incidents, exceeding 3,000 over the 23-year period.
Key Insight: Crash clusters align with major arterial roads and commercial districts, confirming that traffic volume and road type significantly influence crash frequency.

If the map does not load, open directly: Open interactive map

Findings: Most crashes occurred in Clear Weather, more than ten times higher than in rain, and hundreds of times higher than in other conditions.
Key Insight: While clear weather has the highest absolute crash count, adverse weather conditions may have higher crash rates per vehicle-mile traveled.

Findings: A large majority of crashes occur during **Daylight**.
Key Insight: Despite lower traffic volume at night, crashes in darkness show higher severity rates, suggesting reduced visibility significantly impacts crash outcomes.

Reporting Bias: The data only includes crashes that were reported to authorities.
Correlation vs. Causation: This analysis identifies correlations but does not prove causation.
Traffic Volume Normalization: Without traffic volume data, we cannot determine if high-crash areas are truly more dangerous or simply have more traffic.

Based on the findings, future analysis could explore:

Crash Rate Analysis: Normalize crash counts by traffic volume or road length to find truly dangerous areas.
Severity Modeling: Build a predictive model to identify the factors most indicative of a severe or fatal crash.
Time Series Forecasting: Use historical yearly/monthly data to forecast future crash counts.