Interactive road safety dashboard: Integrating spatial analysis and public data through KNIME, QGIS, and Tableau

Traditional road safety analysis relies on static annual reports and absolute accident counts, limiting stakeholders' ability to perform dynamic risk identification across different temporal periods and evaluation perspectives. This study addresses these limitations by developing an interactive dashboard for comprehensive road safety analysis through integration of multiple public datasets using KNIME, QGIS, and Tableau Public. Applied to a part of the Central Bohemian region, a critical area characterized by high traffic volumes and severe accident statistics due to its location surrounding Prague, the framework demonstrates how fragmented data sources can be synthesized into an accessible analytical tool. The methodology encompasses a multi-stage data processing approach where raw car accident data and traffic intensity data undergo pre-processing in KNIME, a spatial analysis phase in QGIS integrates OpenStreetMap data with the official traffic-census network, and the final visualization in Tableau Public offers multiple linked views through four evaluation metrics - the absolute accident count, the Relative Accident Rate (RAR), the Relative Safety Index (RSI), and the Integral indicator of Total Societal Losses (H). Of the 9,325 accidents recorded in the Central Bohemian region in 2023, 4,408 fall on the 643 classified road segments analyzed (1,062 of them with personal injury) ; the analysis reveals that high-risk segment identification varies substantially with the chosen metric - a finding obscured in conventional single-metric reports. Methodological robustness is established through three components: (i) a buffer-size sensitivity analysis (10, 15, 20 m) setting the 15 m matching radius to the 95th percentile of accident displacement, with per-segment results insensitive to the choice (ρ ≥ 0.97); (ii) a Spearman rank-correlation analysis of the metrics (with bootstrap confidence intervals and Kendall's W), showing that they span complementary dimensions while the two severity indicators (RSI and H) capture the same dimension; and (iii) a reproducible TOPSIS multi-criteria ranking of hazardous segments that avoids double-counting the correlated severity indicators (RSI and H). By utilizing open-source tools and public data, the framework provides a potentially scalable, cost-effective solution for road safety management, which may be transferable to other regions and extensible to additional indicators, subject to future practical validation.