Road traffic accidents continue to be a major concern for sustainable urban safety and effective transportation management. Identifying crash hotspots with precision is essential for implementing focused safety interventions. In this study, a severity-weighted system was applied to evaluate crash hazard levels. A Geographic Information System (GIS)-based approach was adopted to analyze spatial patterns of road accidents in Hyderabad, with the primary objective of identifying high-risk locations. Using secondary data from 2021 to 2024, the research employed methods such as Kernel Density Estimation (KDE), Crash Hazard Level (CHL), and Predictive Accuracy Index (PAI) to examine accident frequency, severity, and spatial distribution. A total of 8,576 accident cases were analysed, classified according to factors such as time, location, accident type, and severity. KDE enabled the visualization of accident-prone areas, while CHL and PAI provided a quantitative framework for ranking hazardous zones and validating hotspot predictions. The analysis revealed an increasing trend in non-fatal accidents and highlighted traffic congestion as a major challenge for urban safety. Based on these findings, the study recommends targeted measures, including intersection redesign, enhanced road lighting, pedestrian safety improvements, awareness programs, and better emergency response systems. Overall, the GIS-based approach delivers valuable insights to support urban planners and policymakers in formulating data-driven strategies aimed at improving road safety in Hyderabad.
Introduction
Hyderabad, a rapidly urbanizing city, faces serious road safety challenges due to the interplay of engineering limitations, human behavior, and geographical factors. To address these, the study employs Geographic Information Systems (GIS) integrated with Kernel Density Estimation (KDE) and Crash Hazard Level (CHL) analysis to detect and evaluate accident-prone areas using traffic accident data from 2021 to 2024.
Objectives
Identify major factors contributing to road accident severity in Hyderabad.
Use Kernel Density Estimation (KDE) to locate accident hotspots.
Evaluate crash severity using Crash Hazard Level (CHL) and Severity Index (SI).
Calculate Predictive Accuracy Index (PAI) to validate hotspot detection.
Develop thematic maps for better visualization of accident-prone areas.
Methodology
Data Types:
Non-spatial data: Date, time, vehicle type, injuries, gender, location.
Spatial data: Geocoded accident locations.
GIS Techniques:
Geo-referencing & Digitization: Converting maps into analyzable digital formats.
Layering: Roads, administrative zones, and accidents shown in separate layers.
KDE:
Generates smooth density surfaces to identify accident clusters.
Uses Gaussian kernels with optimal bandwidth (h) for meaningful smoothing.
Crash Hazard Level (CHL):
Severity Index (SI) = L + 3S + 5D
(L: Light injury, S: Serious injury, D: Deaths)
Key Findings
A. Accident Trends (2021–2024)
Year
Fatal
Non-fatal
Property Damage
Total
2021
187
1461
132
1789
2022
143
1158
131
1432
2023
231
1876
285
2392
2024
219
2327
417
2963
Total
780
6822
964
8566
32% increase in total accidents from 2021 to 2024.
Mondays saw the highest number of crashes.
Two-wheelers and pedestrians were the most affected groups.
Men were significantly more involved than women.
Oct–Dec recorded the most accidents.
B. Crash Hazard Level (CHL)
Musheerabad and Nampally emerged as the most dangerous areas, with high KDE values for deaths and injuries.
Severity Index (SI) helped quantify crash risks by combining fatal, serious, and minor accidents into a single score.
C. Hotspot Analysis – East Zone (KDE Results)
Area
KDE Deaths
KDE Injuries
KDE Property Damage
KDE Total
Musheerabad
229.66
259.62
235.22
78.49
Amberpet
195.53
222.22
141.13
62.72
Nampally
195.53
287.03
188.17
78.49
Himayatnagar
162.39
156.81
94.08
31.36
Musheerabad has the highest KDE scores in all categories, making it the top-priority intervention zone.
Nampally has the highest injuries KDE, indicating dangerous driving environments.
Amberpet shows high crash densities, while Himayatnagar is relatively safer but still requires attention.
Conclusion
The accident rate in Hyderabad consistently increased between 2021 and 2024. During this four-year period, the Department of Traffic Police recorded a total of 8,576 accidents. Among these cases, 12% were fatal, 77% were non-fatal, and 10% resulted in property damage only. Overall, the accident rate incresed by 32% from 2021 to 2024, indicating a notable upward trend in road accidents.Kernel Density Estimation (KDE) successfully mapped accident-prone locations, showing high density at Bahadurpura (311.82) and low density at Tirumalgiri (110.36) for fatalities, high density at Secunderabad (290.06) and low density at Bandlaguda (97.44) for injury-related crashes, and high density at Golconda (293.22) and low density at Maredpally (73.45) for property-damage-only cases.Crash Hazard Level (CHL) provided a severity high ranking of accident area is Amberpet 3,796 and low ranking is Shaikpet 79, while Prediction Accuracy Index (PAI) validated hotspot predictions.Results highlighted that non-fatal accidents are 6,822 most frequent, but fatal are 780 and severe crashes remain a significant concern.The core challenge remains the urban traffic congestion and inability to widen the roads at Bandlaguda Charminar and Khairatabad due to dense built-up areas, making targeted interventions the best approach.Spatial models such as GIS and KDE proved vital in identifying patterns and making data-driven decisions for urban traffic safety
References
[1] Bíl, M., Andrášik, R., & Janoška, Z. (2013). \"Identification of hazardous road locations of traffic accidents by means of kernel density estimation environment on road unsafety: Logistic modeling with spatial autocorrelation.\" Accident Analysis & Prevention, 36(6), 1055-1066.
[2] Anderson, T. K. (2009). \"Kernel density estimation and K-means clustering to profile road accident hotspots.\" Accident Analysis & Prevention, 41(3), 359-364.
[3] Erdogan, S. (2009). \"Exploratory spatial analysis of traffic accident statistics and road mortality among the provinces of Turkey.\" Journal of Safety Research, 40(5), 341-351
[4] Plug, C., Xia, J. C., & Caulfield, C. (2011). \"Spatial and temporal visualisation techniques for crash analysis.\" Accident Analysis & Prevention, 43(6), 1937-1946
[5] Wang, C., Quddus, M. A., & Ison, S. G. (2013). \"Impact of traffic congestion on road accidents: A spatial analysis of the M25 motorway in England.\" Accident Analysis & Prevention, 51, 161-171
[6] Li, L., Zhu, L., & Sui, D. Z. (2007). \"A GIS-based Bayesian approach for analyzing spatial-temporal patterns of intra-city motor vehicle crashes.\" Journal of Transport Geography, 15(4), 274-285
[7] Steenberghen, T., Dufays, T., Thomas, I., & Flahaut, B. (2004). \"Intra-urban location and clustering of road accidents using GIS: a Belgian example.\" International Journal of Geographical Information Science, 18(2), 169-181..\" Accident Analysis & Prevention, 55, 265-273
[8] Loo, B. P. (2006). \"Validating crash locations for quantitative spatial analysis: A GIS-based approach.\" Accident Analysis & Prevention, 38(5), 879-886
[9] Blankenship & Stevanovic (2017). \"An automatic car counting system using overfeat framework\"
[10] Savolainen, P. T., & Farzaneh, M. (2011). \"Mixed logit analysis of bicyclist injury severity resulting from motor vehicle crashes at intersection