International Journal of Advanced Engineering Research and

Science (IJAERS)
Peer-Reviewed Journal
ISSN: 2349-6495(P) | 2456-1908(O)
Vol-10, Issue-10; Oct, 2023
Journal Home Page Available: https://ijaers.com/
Article DOI:https://dx.doi.org/10.22161/ijaers.1010.7

Fire Risk Assessment and Emergency Route Decision

Analysis Based on Big Data Platform—Example of
Huizhou
Haolong Liu1, Jingpeng Li2, Yun-Shang Wang3, Ruei-Yuan Wang*4

1
Aba Tibetan and Qiang Autonomous Prefecture Fire and Rescue Detachment
2
Guangdong Zhixiang Technology Co., Ltd
3
Graduate Institute, Fu Jen Catholic University
4
School of Science, Guangdong University of Petrochem Technology (GDUPT), China
*Corresponding author

Received: 16 Aug 2023, Abstract — With the process of urbanization, the frequency of fires is
Receive in revised form: 22 Sep 2023, gradually increasing, and fire emergency work has become particularly
Accepted: 01 Oct 2023, important. This paper takes Huizhou City as an example and conducts a
Available online: 08 Oct 2023 fire risk assessment based on the various conditions of the city using the
©2023 The Author(s). Published by AI Gaode big data platform. The analysis shows that the low-risk area
Publication. This is an open access article within the Fourth Ring Road in Huicheng District is about 91.8%, the
under the CC BY license medium-risk area is about 6.72%, the secondary high-risk area is about
(https://creativecommons.org/licenses/by/4.0/). 1.16%, and the high-risk area is about 0.31%. Meanwhile, we use
Keywords — Gaode Big Data Platform; Fire Python code to calculate the decision analysis plan for fire emergency
Risk Assessment; Emergency Route Decision route selection. By collecting data on fire site locations, fire stations,
Analysis; Analytic Hierarchy Process (AHP); and hospitals, combined with dynamic information such as road
Python conditions and traffic flow, and utilizing the functions of the Gaode big
data platform, route calculation and optimal path selection are carried
out. This method will improve the efficiency of fire emergency response,
reduce rescue time, and provide better assistance to the affected
population.

I. INTRODUCTION process of urbanization, the large gathering of population

Fire is a common natural disaster that poses a huge and buildings, and the increasing complexity of various
threat to the safety of people's lives and property. With the conditions, the risk rate of urban fires is also gradually

www.ijaers.com Page | 70
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

increasing. At the same time, it also increases the number into three categories: qualitative, semi-quantitative, and
of fire risk factors and the difficulty of fire prevention and quantitative (Zhang et al., 2006; Wei, 2019). At present,
control (Hao et al., 2023). For example, in 2021, China most cities in China use the qualitative evaluation method
reported approximately 7,480,000 fires, with 1987 deaths of "urban land classification and fire risk zoning" for fire
and 2225 injuries, resulting in direct property damage risk assessment (Zhang, 2020), and some cities use the
exceeding 6.75 billion yuan (NFRA online, 2022). This indicators selected by the evaluation index system research
threat is very serious in densely populated areas such as institute. Its purpose is to determine the relative fire risk in
commercial and residential communities, so it is necessary a city or a certain area of the city, establish the distribution
to assess and control fire risks and improve urban fire of fires in the city area, and assist the government and fire
safety. In order to maintain the fire safety of the city and departments in analyzing the current situation of fires.
reduce the harm of fire hazards, it is necessary to have an Meanwhile, it is to determine fire risks and acceptable
urban fire risk assessment and emergency strategy levels, allocate fire rescue forces reasonably based on fire
preparation. The former belongs to pre-assessment and safety concerns and cost-benefit analysis, and guide urban
preparation, and early understanding of the areas where fire system transformation and urban fire planning,
risks may occur is necessary to plan and construct various ultimately reducing the level of fire risk in urban areas
basic disaster prevention facilities in order to understand (Zhang et al., 2006). The technology of fire risk
where risks exist, what kind of risks, how to respond, and assessment in urban areas will promote more scientific and
how to improve early management thinking. The latter reasonable fire decision-making.
considers how to deal with, evaluate, and prepare The traditional fire risk assessment mainly includes
emergency measures in the event of a fire. Both thinking establishing an evaluation index system, determining
styles are indispensable and crucial for the indicator weights, and determining the fire risk level (Fu,
decision-making and analysis of fire emergency measures. 2021). Nevertheless, it involves many factors, and the
Firstly, fire risk assessment is a fundamental means of sources of acquisition are complex, making it very difficult
identifying, analyzing, and evaluating the magnitude of to accurately present its specific benefits. It is also often
fire risks through scientific methods to reduce and control influenced by subjective and objective factors (Ketsakorn
fire consequences. It is an important component of fire and Phangchandha, 2023). Currently, research on urban
science research (Fu, 2021) and a fundamental fire risk assessment often focuses on identifying key
decision-making tool in the field of risk control and safety factors and evaluating models. The shortcomings of
management. It provides potential risk information before existing research are mainly reflected in data collection,
a fire event occurs (Gao et al., 2023). Urban fire risk indicator selection, and model selection (Hao et al., 2023).
assessment is an important research topic that supports the With the development of big data platforms and AI cloud
scientific formulation of urban fire safety special planning computing, timely and dynamic evaluation methodologies
and the refined management of fire safety. Urban fire risk are bound to usher in more cost-effective development
assessment is an evolution of fire risk assessment in the prospects. This article attempts to collect Point of Interest
fire protection profession, officially proposed by the (POI) data on fire risk sources, urban characteristics, and
national standard "Urban Fire Planning Specification" and social prevention and control through Python, based on the
identified as an important component of urban fire background of Gaode big data. And combine the weight
planning. The evaluation methods can generally be divided attribute value grading of YAHHP software to establish a

www.ijaers.com Page | 71
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

risk assessment that is different from traditional methods. timely and effective assistance to the affected population.
At the same time, use GIS visualization tools to present the
level assessment classification and distribution of urban II. STUDY AREA AND DATA SOURCES
fire risk. 2.1 Study Area
Secondly, in the management of urban fires, another Huicheng District, under the jurisdiction of Huizhou
important issue is how to quickly and accurately choose City, Guangdong Province, is located in the middle and
the best emergency path after a fire occurs, effectively and lower reaches of the Dongjiang River and the eastern part
quickly eliminate the fire, and carry out personnel rescue, of the Pearl River Delta, with a total area of 1170.6 square
which is an important link to ensuring the safety of kilometers. The development intensity is low, especially in
people's lives and property. With the rapid development of the northern mountainous areas, which are basically
big data technology, the decision analysis of fire undeveloped. It is jurisdiction over 10 streets, 8 towns, and
emergency route selection has gradually attracted the 3 township-level units. As of November 2020, the
attention of academia and engineering practice. permanent population of Huicheng District was 1.559
In previous studies, the selection of fire emergency million. The population density in 2021 is 1347
routes mainly relied on experience and manual judgment, people/square kilometer, ranking at the middle level of the
which have certain limitations. With the rapid development Greater Bay Area (26th). Generally speaking, the
of big data technology, the rich data resources provided by population density of livable cities is generally 1000 to
the Gaode big data platform can be utilized, combined 3000 people per square kilometer, with over 4000 people
with decision analysis methods, to solve the problems belonging to high-density cities. It can be seen that
encountered in fire emergency route selection. It is an Huicheng District is a livable city with an average annual
innovative combination of technology and thinking. The temperature of 22.5 °C, which is consistent with the same
Gaode big data platform has timely and dynamic data on period of the year (1991–2020). The annual extreme
road information, traffic flow, pedestrian flow, and other minimum temperature of 4.5 °C occurred on February 21st.
information across the country. The analysis and mining of The annual extreme maximum temperature is 38.7 °C.
these data will provide more accurate and realistic decision Huicheng District has a good transportation location
support. advantage. To the north is Heyuan City, and to the
Based on the above perspectives, the main purpose of southwest are Dongguan and Shenzhen City. It is the hub
this study is to explore how to utilize the diverse, timely, gateway of the Greater Bay Area, connecting eastern
dynamic, and abundant data sources provided by big data Guangdong, northern Guangdong, and the Fujian Jiangxi
platforms. Through Python programming tools, the region. Within the jurisdiction, there are 6 expressways, 2
decision analysis method for fire emergency route high-speed railways, 2 urban rails, and the Beijing
selection, the establishment of fire risk assessment, and the Kowloon Railway passing through the area. Pingtan
level distribution range were explored. Taking Huicheng Airport and Huizhou Port are within easy reach, and the
District in Huizhou as an example, it is hoped that through three-dimensional transportation network of sea, land, and
the application of this method, accurate risk management air is accessible in all directions. It takes 30 minutes to
mechanisms can be established in daily life. When directly reach Guangzhou and Shenzhen, fully integrating
disasters occur, it can effectively improve the efficiency of into the one-hour living circle of the Greater Bay Area.
fire emergency rescue and reduce rescue time, providing Compared with other counties and districts in the Greater

www.ijaers.com Page | 72
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

Bay Area, it has the position and advantages of an of infrastructure investment in fixed asset investments will
important transportation hub in the eastern part of the rise from 24.83% to 28.13% in 2022. The city's energy
Greater Bay Area. level is constantly rising. It is high time to give full play to
The tertiary industrial structure of Huicheng District the advantages of the central city, gather talents from all
is 3.4: 29.4: 67.1. In recent years, it has continuously sides, and build a regional consumption center.
promoted industrial development and continuously The analysis scope of this study is within the Fourth
optimized the industrial structure, and the commercial and Ring Road area of Huicheng District, Huizhou City
trade industries are large but not strong. The proportion of (enclosed by the Guanglong Expressway, Jiguang
modern service industries is not high, and traditional Expressway, and Changsheng Expressway) (Figure 1),
service industries such as wholesale and retail are still the which is an important dense area and main activity area of
main ones. The level of public service facility investment the city. It also shows that its potential risks are significant
and construction is constantly improving. The proportion and worthy of attention.

Fig.1 Study Area Scope Map of Huicheng District, Huizhou

III. MATERIALS AND METHODS characteristics encompassing population density and fire
3.1 Data Acquisition rescue efficiency; and social prevention and control
This study called the request module through Python encompassing medical institutions..
3.8 to collect POI information data from the Gaode big 3.2 Methods
data platform. Based on the research purpose, obtain POI The concept of this study is mainly based on the
data from three major categories: fire risk sources, urban application of the Gaode big data platform. Firstly, the
characteristics, and social prevention and control. Among AHP method is used to construct a hierarchical structure
them are: fire risk sources encompassing large gas and assign weights using Yaahp software in order to obtain
facilities, gas stations, and flammability; urban the fire risk level distribution map of Huizhou Huicheng

www.ijaers.com Page | 73
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

District and establish a decision-making and control The AHP decomposes decision-making problems into
knowledge map for urban fire management. Secondly, the different hierarchical structures in the order of overall
decision support route selection scheme for emergency objectives, sub-objectives at each level, evaluation criteria,
rescue resources and handling in case of fire is carried out and specific alternative investment plans. Then, by solving
through Python algorithms. By coupling the two, a the eigenvectors of the judgment matrix, the priority
decision-making framework for the urban fire emergency weights of each element at each level to a certain element
decision-making system can be established (Figure 2). at the previous level are obtained. Finally, the weighted
sum method is used to gradually merge the final weights
of each alternative plan into the overall objective. The one
with the highest final weight is the optimal solution
(Figure 3).
This study uses Yaahp (Yet Another AHP) software
for the processing and construction of the analytical
hierarchy process. It is AHP software designed by Beijing
Xinshengyun Software Technology Co., Ltd. This software
provides convenient functions such as hierarchical model
construction, judgment matrix data input, sorting weight
calculation, and calculation data export. The design goal is
that users only need to have preliminary knowledge of the
AHP and do not need to understand various details of AHP
calculations in order to make decisions using the AHP.
Using Yaahp to assist decision-making, users only need to
Fig.2 The Decision-making Framework of This Study pay attention to the decision-making problem itself. With
the model construction function provided by the software,
3.2.1. Analytic Hierarchy Process (AHP) they can organize their ideas and complete a hierarchical
Analytic Hierarchy Process (AHP) refers to a model of the decision-making problem. Then, Yaahp is
decision-making method that decomposes elements that used to fill in the pairwise comparison data in the
are always related to decision-making into levels such as judgment matrix. The software automatically calculates
goals, criteria, and solutions and conducts qualitative and the ranking weights of various decision-making schemes
quantitative analysis on this basis. This method is a based on the hierarchical model and judgment matrix.
hierarchical weight decision analysis method proposed by 3.2.2 Python
American operations researcher Professor Saaty from the Python was designed in the early 1990s by Gido van
University of Pittsburgh in the early 1970s, applying Rossum of the National Center for Mathematics and
network system theory and multi-objective comprehensive Computer Science in the Netherlands as an alternative to
evaluation methods to the study of "power distribution the language ABC. Python provides efficient, advanced
based on the contribution of various industrial sectors to data structures as well as simple and effective
national welfare" for the US Department of Defense object-oriented programming. Python syntax and dynamic
(Alrawad et al., 2023). typing, as well as the nature of interpretive languages,

www.ijaers.com Page | 74
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

make it the programming language for writing scripts and as web development, web crawlers, data science,
rapidly developing applications on most platforms. automated operations and maintenance, database
programming, network programming, graphic processing,
mathematical processing, text processing, multimedia
applications, artificial intelligence, and automated testing.
Among them are the Internet of Things and robot control.
Web development can be combined with Python, HTML,
CSS, JavaScript, databases, and other applications to
develop a website. In addition, data science includes
machine learning, data analysis, and data visualization. In
addition, Python can be used for machine learning, which
can research artificial intelligence (AI), robotics, language
recognition, image recognition, natural language
processing, expert systems (ES), as well as data analysis
and visualization.
This study is based on the data source of the big data
Fig.3 Process Flow of Processing Program Judgment by platform, using Python as a tool to form a decision support
AHP rule program, combined with dynamic and timely data, to
form a judgment decision plan, and using GIS as a visual
What can Python do? With the advent of the AI era, display tool to generate emergency decision-making (Li
the Python language can provide diverse applications such and Wang, 2023) (Figure 4).

Fig.4 Python and GIS base processing

3.2.3 Algorithms Construction on the decision support concept of event occurrence, to

This study uses the Python language as a tool, based simulate and think about the handling methods when a fire

www.ijaers.com Page | 75
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

occurs. Firstly, based on the data content of big data, possible for rescue. If a wounded person needs to be
search for nearby fire stations and calculate the distance, rescued, the search and condition calculation of the
time, road conditions, traffic flow, traffic lights, and other hospital site must also consider various dynamic
related conditions between the stations and the event conditions of path selection to improve the efficiency of
location to find the optimal fire station and time path. rescue (Figure 5).
Provide a fire brigade to arrive at the scene as soon as

Fig.5 Thinking Process for Emergency Decision-making

IV. ANALYSIS AND RESULTS fire rescue efficiency, and medical institutions) (Figure 6),
4.1 Fire Risk Level Assessment and establish weight values (Figure 7), which are 0.178,
4.1.1 AHP Analysis 0.022, 0.052, 0.106, 0.2, and 0.041, respectively. And
This study use Yaahp software to evaluate and grade using GIS buffer analysis, point density analysis, and
the POI data of three major categories, including fire risk grading based on element weight attribute values, the
sources, urban characteristics, and social prevention and impact of each element on fire risk assessment can be
control, as well as six subcategory indicators (such as large visually displayed in the grid through color separation
gas facilities, gas stations, flammability, population density, (Figure 8).

www.ijaers.com Page | 76
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

Fig.6 Index Evaluations for AHP Analysis

Fig.7 Indicator Weight Table by Yaahp

www.ijaers.com Page | 77
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

Fig.8: Evaluation and Grading Status of Six Subclass Indicators

Note: F-2, F-2, F3 are Fire Risk Sources; U-1, U2 are Huicheng District are divided into four levels to
Urban Characteristics; S-1 is Social Prevention and demonstrate their spatial distribution patterns (Figure 9).
Control According to statistics, about 91.8% of the low-risk
4.1.2 Risk Level Classification and Visualization area, 6.72% of the medium-risk area, 1.16% of the
After overlaying the above three-level indicator data, secondary high-risk area, and 0.31% of the high-risk area
ArcGIS software is used to classify the evaluation results are within the Fourth Ring Road in Huicheng District.
into four levels: low-risk, medium-risk, secondary Based on the visualized spatial distribution pattern, the
high-risk, and high-risk. The evaluation results are secondary high-risk areas and high-risk areas are mainly
combined with GIS for visual display. Finally, based on concentrated in Henan-An Street, and there are also some
the comprehensive calculation results of fire risks in each secondary high-risk areas in Jiangbei Street and Shuikou
sub-region, the fire risks of each evaluation grid unit in Street.

www.ijaers.com Page | 78
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

.
Fig.9 Distributions of Fire Risk Levels in Huicheng District

4.2 Emergency Decision Analysis of Events sent to the nearest Xiapu People's Hospital in Huizhou
In this case, assuming a fire occurs in a high-risk area City.
of Henan-An Street, based on the calculation rule of Based on the above, the area is located in a high-risk
decision support, the neighboring fire stations are area, and the required rescue time during normal and peak
determined (Figure 10). The figure shows four nearby fire hours is also different. During peak hours, the traffic flow
stations. Through dynamic and timely data from the big will be relatively high, and fire trucks will run slowly. This
data platform, the distance, time, road conditions, traffic is a dynamic difference in the spatio-temporal accessibility
flow, traffic lights, and other related conditions between of transportation, not a fixed value. Through dynamic big
the locations are calculated. The time is determined by data calculations, the timeliness and accuracy obtained will
“distance * speed” and traffic lights are comprehensively be more accurate than traditional decision-making based
considered. The results obtained from timely data solely on road length. So, although it appears that the
calculations, such as traffic factors, the calculation in this support fire station for rescue is far from the fire scene in
case shows that the nearest point is the fire brigade on the this example, after comprehensive calculation of various
south bank of the river. It can quickly reach the fire site for data conditions, it shows that its rescue service time is the
rescue in about 5 minutes through Yanda Avenue and be shortest.

www.ijaers.com Page | 79
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

Fig.10: Case study of optimal path selection for fire emergency rescue

V. CONCLUSIONS data sources to form more effective decision plans through

This study mainly focuses on the analysis of urban the true form of data in order to approach the actual
fire risk and emergency decision-making management. situation more closely and achieve the timeliness and
Based on the dynamic data of big data platforms, AHP effectiveness of decision-making. Through the analysis
evaluation is used to establish the level of urban fire risk, framework proposed by this study, the concept and
which serves as the basis for decision-making conditions analytical benefits of emergency decision support are
for fire backup management in daily life. And use the demonstrated, which can enhance the actual effectiveness
Python language for data integration, calculation, and of fire prevention and provide reference for future relevant
dynamic analysis. Firstly, search for nearby fire stations departments to establish relevant decision-making
and hospitals based on the location of the fire and calculate systems.
the optimal route to improve the efficiency of fire
emergency response. Secondly, considering the road REFERENCES
conditions, traffic flow, and other information in real [1] Alrawad, M., Lutfi, A., Almaiah, M.A., Alsyouf, A., Arafa,

situations, utilizing the functions of the Gaode big data H.M., Soliman, Y., and Elshaer, I.A. A Novel Framework of

platform for route calculation and selection ensures that Public Risk Assessment Using an Integrated Approach

rescue operations can quickly reach the fire scene and Based on AHP and Psychometric Paradigm.Sustainability.

escort the injured to the hospital. 2023; 15(13):9965. https://doi.org/10.3390/su15139965

In the era of big data and AI, all decision analysis and [2] Fu, X.Q. A review of fire risk assessment methods in urban

calculations should rely more on timely and dynamic big area, Fire Science and Technology, 2021, 40(1), 1622-1625

www.ijaers.com Page | 80
Liu et al. International Journal of Advanced Engineering Research and Science, 10(10)-2023

[3] Gao, S., Huang, G, Xiang, Z., Yang, Y., Gao, X. A Novel

Risk Assessment for Cable Fires Based on a Hybrid

Cloud-Model-Enabled Dynamic Bayesian Network Method.

Applied Sciences. 2023; 13(18): 10384.

https://doi.org/10.3390/app131810384

[4] Hao, Y., Li, M., Wang, J., Li, X., and Chen, J. A

High-Resolution Spatial Distribution-Based Integration

Machine Learning Algorithm for Urban Fire Risk

Assessment: A Case Study in Chengdu, China. ISPRS

International Journal of Geo-Information. 2023; 12(10):404.

https://doi.org/10.3390/ijgi12100404

[5] Ketsakorn, A., and Phangchandha, R. Application of

Analytic Hierarchy Process to Rank Fire Safety Factors for

Assessing the Fire Probabilistic Risk in School for the Blind

Building: A Case Study in Thailand. Fire. 2023; 6(9):354.

https://doi.org/10.3390/fire6090354

[6] Li, J. & Wang, R.Y. Analysis of Spatial-Temporal

Accessibility in Huizhou West Lake Scenic Spot Using Big

Data for Smart City, Global Journal of Arts Humanity and

Social Sciences, 2023, 3 (2), 165-172.

[7] National Fire and Rescue Administration (NFRA).

Firefighters Responded to a Record Number of Calls in

2021, Fighting 745,000 Fires. 2022, Available online:

https://www.119.gov.cn/gk/sjtj/2022/26442.shtml

[8] Wei, Y. Quantitative Assessment Method and Application

Research on Urban Regional Fire Risk, Wuhan University

of Science and Technology, 2019

[9] Zhang, L. Based on GIS Fire Risk Assessment-Taking

Lishui District, Nanjing City as an Example. Science and

Technology & Innovation, 2020, 24:48-52. DOI ：

10.15913/j.cnki.kjycx.2020.24.015

[10] Zhang, X., Wang, B., Du, X., Li, J., and Wu, C. Exploration

on the method of regional fire risk assessment in city,. Fire

Science and Technology, 2006, 25(2), 198-201.

www.ijaers.com Page | 81