An interactive Power BI dashboard that analyzes road accident and casualty data across the UK, uncovering trends by vehicle type, road type, weather, location, and time — helping stakeholders identify high-risk conditions and prioritize road safety interventions.
Road accidents remain a critical public safety concern, and understanding where, when, and under what conditions they occur is key to reducing casualties. This project transforms raw accident records into a single-page, decision-ready Power BI report that lets users slice data by road surface and weather conditions, and instantly see the impact on casualties across multiple dimensions.
The dashboard compares Current Year (CY) performance against the Previous Year (PY), surfacing year-over-year trends and highlighting whether casualty numbers are improving or worsening.
- Track overall casualty and accident volume trends year-over-year
- Break down casualty severity (Fatal, Serious, Slight)
- Identify which vehicle types are most involved in accidents
- Compare accident patterns across road types, light conditions, and urban vs. rural areas
- Visualize the geographic distribution of accidents across the UK
- Enable dynamic filtering by road surface and weather conditions for root-cause analysis
| Metric | Description |
|---|---|
| Total CY Casualties | 51.2K (-17.1% vs PY) |
| Total CY Accidents | Year-to-date accident count (-16.8% vs PY) |
| CY Fatal Casualties | 1.1K (-40.6% vs PY) |
| CY Serious Casualties | 8,041 (-22.7% vs PY) |
| CY Slight Casualties | 42.1K (-15.0% vs PY) |
- Casualties by Vehicle Type — breakdown across Agricultural, Bike, Bus, Car, Other, and Van
- CY vs PY Casualties Monthly Trend — area chart comparing 2021 vs 2022 casualty trends across all 12 months
- Casualties by Urban/Rural — donut chart split (Urban 62.47% / Rural 37.53%)
- Casualties by Road Type — bar chart across Single Carriageway, Dual Carriageway, Roundabout, One Way Street, and Slip Road
- Casualties by Light Conditions — donut chart split (Day 73.84% / Dark 26.16%)
- Casualties by Location — interactive map plotting accident locations across the UK
- Road Surface — filter all visuals by surface condition
- Weather Conditions — filter all visuals by weather at time of accident
- Power BI Desktop — data modeling, DAX measures, and report design
- Power Query (M) — data cleaning and transformation
- DAX — YoY comparisons and dynamic KPI calculations
- Bing Maps visual — geographic accident plotting
| File | Description |
|---|---|
Road_Accident_Analysis.pbix |
Power BI project file containing the full data model, DAX measures, and report |
Road_Accident_Dashboard.png |
Static preview/screenshot of the dashboard |
README.md |
Project documentation |
- Clone or download this repository.
- Ensure you have Power BI Desktop installed (free).
- Open
Road_Accident_Analysis.pbix. - Use the Road Surface and Weather Conditions slicers in the top-right to explore the data interactively.
- Casualties dropped 17.1% year-over-year, with fatal casualties down 40.6% — the largest improvement across all severity categories.
- The majority of casualties occur in urban areas (62.47%) and during daylight hours (73.84%), suggesting that traffic volume/density is a stronger factor than visibility.
- Cars account for the overwhelming majority of casualties by vehicle type, followed by vans and bikes.
- Single carriageways are by far the most common road type associated with casualties (~37K), far exceeding dual carriageways and roundabouts.
- Casualty volumes show a clear seasonal pattern, dipping mid-year and rising sharply toward the end of the year (Nov/Dec).
- Add drill-through pages for detailed incident-level analysis
- Incorporate time-of-day analysis alongside light conditions
- Add predictive modeling (e.g., risk scoring by road/weather combination)
- Publish to Power BI Service with scheduled data refresh
Ashmit Srivastava
- 🔗 GitHub: github.com/ashgithub0208
- 💼 LinkedIn: linkedin.com/in/ashmit-srivastava0208
Feel free to connect or reach out with questions, feedback, or collaboration ideas!
This project is open-sourced for educational and portfolio purposes. Feel free to fork and adapt it.