GIS has an important and distinctive role in spatial analysis, so what are the best methods for spatial analysis of road accidents?
Dear Qayssar,
I think there is no 'best' method. There are several different approaches that can answer different questions. Both 'spatial analysis' and 'road accidents' are very broad topics, and the method to be selected depends not only on the research question but on the type, quality and resolution of your data, and also on the presence of auxiliary variables.
Best,
Ákos
Dear Qayssar,
I think there is no 'best' method. There are several different approaches that can answer different questions. Both 'spatial analysis' and 'road accidents' are very broad topics, and the method to be selected depends not only on the research question but on the type, quality and resolution of your data, and also on the presence of auxiliary variables.
Best,
Ákos
Qayssar Ajaj Terrain, location, type of road, curves, bends, speed limits, etc are some of the factors that lead to the cause of accidents. You can carry out all these analyses on the road network spatial data available using any GIS software Viz. Arc GIS; open-source - Q GIS, etc., so that you can pinpoint the accident-prone points.
There are methods, factors, algorithms, tools.
Methods: observation, statistical, visualization, etc.
Factors: listed above.
Algorithms: depends on specific task.
Tools: GIS, GPS, cameras database, etc.
What kind of road exident? You need to specify your objective. Spatial analysis for road exident is too broad.
I f you want the road accident cluster, you can use the hotspot analysis.
Statistically significant hot spots and cold spots in a dataset can be
located using hotspot analysis. The hotspot analysis can be done using the
Getis-Ord (G i*) local statistical technique which produces Z scores and P
values from the dataset. A small P-value and a high Zscore indicates a
significant hot spot while a small P-value and low negative Z score
indicate a significant coldspot. The higher (or lower) the Z score, the
more intense the hotspot (coldspot) clustering. A Z score near zero
means no spatial clustering.
However, you need to specify your objectives.
GIS was used as a management system for accident analysis by applying combination of spatial-statistical methods.
The operational approach of spatial patterns was developed in geographical information system (GIS) framework to analyze three types of urban accidents (fatal, injury and property damage only (PDO)).
https://reader.elsevier.com/reader/sd/pii/S2095756417301988?token=2AD4A465F2E9EC5B4242ADD9EA640DD46479D85ED12ADF736A17F760E05F89E79557A581AC75F1DE847C52D8CC72B30F
A GIS based traffic accident data collection, referencing and analysis framework for Abu Dhabi
As long as I known.. I suggesting you to use method Kernel Density. I thinking it most better to make spot zone of this accident.
Article GIS-based spatial analysis of urban traffic accidents: Case ...
it will help you.
GIS-based spatial analysis of urban traffic accidents: Case study in Mashhad, Iran
The following conclusions can be drawn for the total number of accidents:
For the number of accidents per unit of time, there is a clear and strong time dependence. The number of accidents is higher in April-may, and less in the winter months. The peak of accidents every day falls on the evening hours, when people return from work. The largest number of accidents occurs at the end of the working week, the smallest-at the weekend.
The accident rate is strongly influenced by systematic factors such as weather conditions, the presence of structures, and the condition of the pavement. Lighting (operation of lighting devices) also affects, and the included headlights reduce accidents even in the twilight/light time.
The most frequently recorded are pedestrian collisions and minor collisions with 1-2 victims. Collisions mostly occur at intersections, and collisions occur at pedestrian crossings.
Accidents on major highways occur less frequently than on medium-sized roads, but are characterized by a large number of victims.
As for the number of victims of accidents, the following conclusions can be drawn:
There are clear and strong time factors depending on the number of victims from the characteristics of the accident, and periods characterized by a decline in accidents have a high average number of victims.
The largest number of victims is characterized by accidents with a large number of participants (people). But a small number of cars.
The number of victims is almost not subject to systematic factors, such as the state of the road or the weather.
When building a model of the dependence of the number of victims on the accident parameters, the most significant factors are the type of incident and time parameters.
Accidents on small roads are characterized by a small number of victims, while on large motorways the number of victims is increasing.
It all depends on your requirement, objectives, type of analysis and the way you are planning to present your outcomes.
Agree with all of the above. Would need an idea of the dataset and research question(s). Is it site specific or a large dataset of collision hotspots? (Related side note - even though it hasn't been completely adopted in the academic vernacular, safety advocates push for using the language of "crashes" or "collisions", rather than "accidents". The latter term implying that vehicle collisions are just oopsy-type collateral damage of automobility, rather than preventable tragedies (for e.g. see: The Automobile Imperative).
للتحليل المكاني عدة طرق حسب شكل ونمط الظاهرة فممكن ترجع لكتاب جمعة داوود الخاص بالتحليل المكاني ستجد فيه انواع التحليل المختلفة ..تحياتي لك
Hotspot identification using: KDE, Emirical Bayes Estimate, DBSCAN, etc.
Collect good spatial and temporal data. When an accident occur, record the location in form of x and y.
This research article and book will be helpful:
https://www.tandfonline.com/doi/pdf/10.1080/jom.2007.9710827
or,
Book Spatial Analysis Methods of Road Traffic Collisions
- Badges
- Science topic
- Similar topics
- Geoscience
- Geoinformatics
- Geographic Information System