Methodology for identifying accident-prone zones based on spatial clustering of road accident data

Introduction. Improving road safety requires the development of methods for proactive identification of areas with an increased concentration of road accidents. Existing approaches do not sufficiently use the potential of analyzing historical spatiotemporal data to predict accident-prone zones.
Materials and methods. The work uses data on 16,247 road accidents for the period from 2015 to 2024 in Kazan, provided by the State Traffic Safety Inspectorate. The method of spatial clustering over a fixed radius (100 m) with a density threshold of 20 accidents per zone has been applied. Python libraries (pandas, haversine) have been used for data processing, folium and GeoJSON for visualization.
Results. Methodology and algorithm for identifying accident-prone zones have been developed, including the stages of preprocessing geodata, clustering by radius, accident density estimation and zone characteristics by dominant types of accidents. 127 accident-prone zones have been identified, 50 most significant are characterized by concentrations from 21 to 87 accidents. 78% of these zones were found to be located at intersections and areas with heavy traffic. A temporal analysis revealed a peak accident rate in the evening (5:00 p.m. – 8:00 p.m.) and an increase in the number of accidents by 23% in the autumn-winter period.
Discussion and conclusion. The proposed methodology makes it possible to identify objectively accident-prone zones found by the clustering algorithm, taking into account the spatial concentration of incidents. The results can be integrated into navigation systems to generate context-sensitive warnings to drivers. The practical significance is confirmed by the development of a prototype mobile application on the Android platform with the use of Yandex MapKit SDK.
Practical value: The technique gives an ability to reduce labor costs for identifying problem sections of the road and increases the efficiency of resource allocation for road maintenance services. A prototype of an Android-based mobile application has been developed with the use of Yandex MapKit SDK, which provides visual and audible warnings to drivers approaching accident-prone zones.
Originality/value: An integrated approach combining spatial analysis of State Traffic Safety Inspectorate data with the ability to visualize quickly and integrate into mobile navigation services for proactive driver warnings.

1. Kurbatov D.S., Starostenko A.V. Innovationnye podhody k organizacii dorozhnogo dvizheniya na osnove cifrovyh dvojnikov avtomobil’nyh dorog. The Russian Automobile and Highway Industry Journal. 2025;22(5):772–785. (in Russ.) DOI: 10.26518/2071-7296-2025-22-5-772-785

2. Trofimenko Yu.V., Shashina E.V. Vliyanie chelovecheskogo faktora na obespechenie bezopasnosti dorozhnogo dvizheniya. Bezopasnost’ zhiznedeyatel’nosti. 2016;1(181):24–27. EDN VIUSCP. (in Russ.)

3. Reason J. Human error: models and management. BMJ. 2000; 320(7237): 768–770. DOI: 10.1136/bmj.320.7237.768

4. Vaa T. Cognition and emotion in driver behaviour models: some critical viewpoints. Institute of Transport Economics. Oslo; 2007. p. 48–59.

5. Ermagun A. Transit access and urban spacetime structure of American cities. Journal of Transport Geography. 2021;93:103066. DOI: 10.1016/j.jtran-geo.2021.103066

6. Szakonyi P., Makó E. Defining Correlation between the Modal Split of Inhabitants and Students and the Location of Housing Areas and Schools with the Analysis of Travel Plans. Transportation Research Procedia. 2014;4:271–285. DOI: 10.1016/j.tr-pro.2014.10.007

7. Il’ina I.E., Vitvickij E.E. Indeksy dlya ocenki urovnya bezopasnosti dorozhnogo dvizheniya v regionah Rossii. The Russian Automobile and Highway Industry Journal. 2025;22(1):68–77. (in Russ.) DOI: 10.26518/2071-7296-2025-22-1-68-77

8. Yakunin N.N., Yakunina N.V., Fattahova A.F., Postnikova A.A., Yanuchkov M.R. Metodika prognozirovaniya transportnoj potrebnosti naseleniya gorodov s vysokim turisticheskim potencialom v pikovyye sezony. Vestnik SibADI. 2024;21(2):256–269. (in Russ.) DOI: 10.26518/2071-7296-2024-21-2-256-269

9. Akchurin A.V., Kondratova Yu.N. Opredeleniye oblastej povyshennoj dorozhno-transportnoj opasnosti na karte goroda s pomoshch’yu klasterizacii dannyh. Matematicheskoe modelirovanie, komp’yuternyj i natural’nyj eksperiment v estestvennyh naukah. 2023;(4):45–52. (in Russ.)

10. Alekseeva A.A., Volchatova I.V. Analiz prostranstvennogo raspredeleniya DTP s ispol’zovaniyem GIS-tehnologij. Bezopasnost’–2025. Problemy tekhnosfernoj bezopasnosti. 2025:79. (in Russ.)

11. Zagidullin R.R. Povysheniye effektivnosti dorozhnogo dvizheniya dinamicheskim upravleniyem skorosti dvizheniya transportnyh sredstv na primere g. Kazani. The Russian Automobile and Highway Industry Journal. 2025;22(6):940–951. (in Russ.) DOI: 10.26518/2071-7296-2025-22-6-940-951

12. Mishra M., Kumar A. ADAS Technology. In: Advanced Driver Intention Anticipation and Ecodriving Assistance Systems. CRC Press; 2022. p. 415–435. DOI: 10.1201/9781003048381-21

13. Lazarev D.A. Mnogokriterial’nyj metod ocenki urovnya bezopasnosti dorozhnogo dvizheniya na uchastke ulichno-dorozhnoj seti. The Russian Automobile and Highway Industry Journal. 2025; 22(6): 952–965. (in Russ.) DOI: 10.26518/2071-7296-2025-22-6-952-965

14. Aleksa M., Schaub A., Erdelean I. et al. Impact analysis of Advanced Driver Assistance Systems (ADAS) regarding road safety – computing reduction potentials. European Transport Research Review. 2024; 16:39. DOI: 10.1186/s12544-024-00654-0

15. Rybin A.L., Kretov V.A. Rezul’taty ocenki effektivnosti meropriyatij po povysheniyu bezopasnosti dorozhnogo dvizheniya metodom vygod i zatrat. Dorogi i mosty. 2020; (1): 46–62. (in Russ.)

16. Lin J.L. Generalizing local density for density-based clustering. Symmetry. 2021; 13(2): 185. DOI: 10.3390/sym13020185

17. Bogdanovich S.V., Kapskij D.V. Analiz vozmozhnostej primeneniya metodov mashinnogo obucheniya i analiza bol’shih dannyh dlya vyyavleniya faktorov riska v upravlenii bezopasnost’yu dorozhnoj infrastruktury. Sovremennye tehnologii v transportnoj otrasli. 2024:114–117. EDN UJTJHC. (in Russ.)

18. Nevolin D.G., Carikov A.A., Sorogin I.G. Sezonnaya neravnomernost’ dorozhno-transportnyh proisshestvij na avtomobil’nyh dorogah Sverdlovskoj oblasti. Avtotraktorostroenie i avtomobil’nyj transport. 2023;2:4–10. EDN SBFNTB. (in Russ.)

19. Zervou M. Python Data Cleaning and Preparation Best Practices. Birmingham: Packt Publishing; 2024. 456 p.

20. Suma K.G., Sunitha G., Avanija J., Galety M.G., Varna C.P. Geospatial Data Visualization With Folium. In: Geospatial Application Development Using Python Programming. IGI Global; 2024. p. 187–208.

21. Giniyatov Ch.R. Kak mashinnoye obucheniye i analiz dannyh menyayut podhod k bezopasnosti dorozhnogo dvizheniya. Obrazovaniye, nauka i innovacii: sovremennyye vyzovy. 2025:115–117. EDN WJIQGK. (in Russ.)

22. Simakova A.E., Syrkin I.S. Ispol’zovaniye API Yandex.Kart. Rossiya molodaya. 2022:3291–3295. (in Russ.)

23. Adnan M., Tayyab M., Ahmad H.M.W., Imran M., Anjum M.N. The impact of road safety education on young drivers’ behavior and attitudes: A systematic review and meta-analysis. Transportation Research Part F: Traffic Psychology and Behaviour. 2025; 109: 94–112. DOI: 10.1016/j.trf.2024.11.025

24. Evtyukov S.S., Brylev I.S., Blinder M.M. Assessment of the impact of the cycling infrastructure of the city on the road safety of cyclists. World of Transport and Technological Machines. 2022; 3-2(78): 76–84. DOI: 10.33979/2073-7432-2022-2(78)-3-76-84

25. Novikov A.N., Eremin S.V., Lomakin D.O. Assessment of road safety at the regional level. World of Transport and Technological Machines. 2020; 3(70): 72–79. DOI: 10.33979/2073-7432-2020-70-3-72-79

26. Sil’yanov V.V., Kapitanov V.T., Monina O.Yu., Chubukov A.B. Prognoz chisla pogibshih v DTP na osnove social’no-ekonomicheskih pokazatelej. Nauka i tehnika v dorozhnoj otrasli. 2017; (4): 4–6. (in Russ.)

27. Surova L.V. Rol’ chelovecheskogo faktora v obespechenii bezopasnosti [The role of the human factor in safety assurance]. Vestnik Kazanskogo gosudarstvennogo energeticheskogo universiteta, 2012; 2(13): 64–71. EDN QZXOZB.