This is the author version of the conference paper in CRIOCM15.
Full version available at: https://link.springer.com/chapter/10.1007/
978-981-10-0855-9_36

The Experiences for Post-disaster
Reconstruction in Rural China
The Implications from Data Mining

Yi Peng, Fan Xue and Haijun Bao

Abstract With rapid economic development and increasing devastation of natural
hazards, rural China has, and would still have, large amount of post-disaster reconstruction. Many reconstruction cases were reported, especially after the 5.12 Sichuan
Earthquake. However, there is a lack of systematic studies on the experiences for
post-disaster reconstruction in Rural China. Therefore, the precious experiences
gained from one reported case are difficult to deliver to other cases facing natural
disasters in future. In order to mitigate this deficiency, this study adopts data mining to
systematically investigate post-disaster reconstruction in rural China. Considering
news reports have a broader audience than that of academic papers, the source of
reconstruction cases is reliable online news reports. The time period for consideration
is from 12 May 2008 to 31 May 2015. The keywords for identifying possible news
reports are “post-disaster reconstruction” and “rural areas”. Preliminary refinement is
conducted to exclude the news reports introducing policies rather than reconstruction
cases. The similarities and differences of the selected reconstruction cases is mined in
terms of manpower, organization, financial source, reconstruction approach (concentration versus in situ), identified experiences, and identified problems. Based on
the results, the experience and problem checklist is developed for better experience
sharing of post-disaster reconstruction in rural China. The results can also facilitates
the government and NGOs to find suitable measures to promote sustainable development in post-disaster reconstruction.
Keywords Post-disaster reconstruction
sharing


 Rural China 
 Data mining 
 Experience

Y. Peng (&) 
 H. Bao
School of Urban-Rural Planning and Management, Zhejiang University
of Finance and Economics, Hangzhou, China
e-mail: pengyihz@gmail.com
F. Xue
Department of Building and Real Estate, The Hong Kong Polytechnic University,
Hung Hom, Hong Kong

1

Introduction

China is a country with a highly frequency of natural disasters. According to the
database of EM-DAT [1], from 1900 to 2011, there were 681 natural disasters, which
resulted in about 12 million deaths, 63 million homelessness, and damages costing
nearly 363.43 billion US dollars. Earthquake is one of the most serious natural disasters in China, which has had 1/3 of the devastating earthquakes worldwide even
though the land area of China occupies only 1/4 of the world [2]. China recorded an
earthquake of magnitude 6.0 or higher 26 times, including earthquake of magnitude
7.0 or higher 7 times [3]. Moreover, it is found that more than 80 % of the earthquakes
above magnitude 5.0 occurred in rural areas and brought huge losses [4]. Natural
disasters bring about large number of deaths, make countless others homeless, and
cause a huge amount of damage in both rural and urban areas. It is difficult to compare
the losses between rural areas and urban areas caused by natural disasters as there is no
international database of disaster losses that disaggregates data by urban and rural area
[5]. However, it is considered that the rural areas face more disadvantages than the
urban areas due to insufficient infrastructure, lack of information about natural disasters, and social inequality. For example, housing damage brought about by the 5.12
Sichuan Earthquake in rural areas was much higher than in urban areas of Sichuan
Province. As well, it is found that the housing damage is the main loss for the farmers
especially in developing countries, such as China [6].
Recent research has focused on post-disaster reconstruction in rural areas of
developing countries. Post-disaster reconstruction is a critical, integral part of
recovery. Post-disaster reconstruction stresses rebuilding of the physical structures,
such as infrastructure and house destroyed or damaged in a natural disaster [7].
Various models have been developed to investigate the reconstruction process
[8–10]. The researchers, especially the social researchers, stress social process is
important in the recovery after a disaster [11]. Therefore, the recent research moved
beyond physical reconstruction, and laid the concern for the social process and
sustainable post-disaster reconstruction [12–14]. However, physical reconstruction
provides the basis for settling down the victims, restarting industry production, hence
facilitating economic recovery, and restoring the environmental function. As physical
reconstruction is performed by people, community restoration or reformation such as
community participation, and a decentralized approach can be better implemented
through means of physical reconstruction [15]. Physical reconstruction is even
important for psychological recovery, as the physical building provides the sense of
place, which is critical and helpful in achieving psychological recovery [16]. As a
result, an increasing number of studies have been conducted on post-disaster
reconstruction, particularly in housing reconstruction and relevant aspects.
However, research on an effective mechanism and system to identify, store, and
distribute the experiences of housing reconstruction is few. As Hegel criticized that
“what experience and history teach is this,-that peoples and governments never have
learned anything from history, or acted on principles deduced from it”, there is a lack
of systematic studies on the experiences for post-disaster reconstruction in Rural China.

In order to mitigate this deficiency, this study aims to systematically investigate
experiences of post-disaster reconstruction in rural China through data mining.
Section 2 introduces the research methods used in this study. Section 3
presents the findings from data mining. Thorough discussions have been conducted
in this section. Section 4 concludes this research with presenting future studies.

2

Research Methods

In order to achieve the research aim, the research flow as shown in Fig. 1 has
been developed.
The established mining aim would guide the whole mining process from defining
the source, boundary, and keywords. In addition, the established aim also confines
the results interpretation. The aim of the data mining in this study is to systematically
investigate the experiences of post-disaster reconstruction in rural China.
There are various types of media recording the experience, such as electronic,
paper, audio and video recording. Considering electronic recording is easily
Fig. 1 The research flow of
this study

Step 1

Specify the aim of mining

Step 2

Define the source, boundary, and
keywords of text mining

Step 3

Check and refine the selected cases

Step 4

Conduct text mining

Step 5

Analyze and interpret the results

accessed for analysis, this paper confines the analysis to electronic recording. In
addition, as news reports have a broader audience than that of academic papers, the
source of reconstruction cases is therefore chosen as reliable online news reports.
The online website of news reports was chosen from the rank of credible websites,
which includes popular websites such as Xinhua Net, Tencent News [17].
Moreover, the time boundary for searching in such sources should be defined
otherwise, there would be unexpected ineffective work. This study defines the time
period for consideration is from 12 May 2008 to 31 May 2015. Keywords is useful
to find the possible cases in the defined sources. The keywords for identifying
possible news reports in this study are “post-disaster reconstruction (or post-disaster
recovery)” and “rural areas”.
By following the first two steps, it is easy to find the possible news reports
through searching. Preliminary refinement will be conducted to exclude the news
reports introducing policies rather than reconstruction cases. This step is necessary
as there are usually some news reports just delivering the formulated policies of
post-disaster reconstruction. In addition, the exactly same news reports occurred in
different websites was only consider once. It should be noticed that different news
reports of the same village is considered as different news reports as different
information would be provided in such condition.
After confirming the news reports, data mining would be conducted to select the
concerned information in the large pool of news reports. As usually text was used to
introduce the reconstruction cases in the electronic news reports, text mining would
be used in this study. Text mining, or text data mining, refers to the process of
deriving high-quality information from text. Typical text mining tasks include text
categorization, text clustering, concept/entity extraction, production of granular
taxonomies, sentiment analysis, document summarization, and entity relation
modeling [18]. Python was used to develop special programs to mine the reconstruction cases.
Special concerns were paid to the similarities and differences of these cases in
terms of manpower, organization, financial source, reconstruction approach (concentration versus in situ), identified experiences, and identified problems.
Interpretation of the findings can help to understand what are emphasized in
recording such experiences of post-disaster reconstruction in rural China. The
understanding of similarities and differences in different reconstruction cases is
useful for experience sharing.

3

Findings and Discussions

3.1 Findings
By following the steps introduced in the Section of research methods, this study
found 835 news reports with the keywords “post-disaster reconstruction

Table 1 Statistical results of the data mining
Concerns

Number of cases

Key points

Manpower

124

Organization

80

Finance

88

Reconstruction
approach

56 (concentration)
27 (relocation)
11 (in situ)
34

Insufficiency of reconstruction professionals,
education after disasters, number of rural victims
Governmental organization improves the
reconstruction results, discussion of organization
modes suitable for reconstruction, contributions of
NGOs, organization process in post-disaster
reconstruction
Investment on reconstruction, financial audit,
financial support from brother provinces
Benefits of concentrated reconstruction, how to
determine the reconstruction approaches, urban-rural
coordinated development
Experiences in terms of organization, site selection,
financial sources, housing allocation
Problems in terms of organization, site selection,
financial sources, housing allocation

Introducing
experiences
Introducing
problems

75

(or post-disaster recovery)” and “rural areas”. After excluding the news reports
introducing policies rather than experiences and occurring in different websites,
there were only 532 useful news reports for analysis. As shown in Table 1, not
Table 2 Top 10 experiences and problems
Rank

Experience

Problems

1

Transparent and consistent policies

2

Many matched policies to support
post-disaster reconstruction
Sufficient promotions of relevant policies

Undetermined property rights of
cultivated land generated from
consolidating former rural
residential land
Incapability of completing project
application before reconstruction
Insufficient financial support on daily
management for CRS
Little attention to psychological
recovery

3
4

5
6
7
8
9
10

Establishing special village affairs board
and special supervision board to monitor
reconstruction
Coordination of governments at different
levels
Assessment on degree of damage rather
than the economic values of the houses
Using local or nearby materials and
construction teams
Following rural victims’ willingness
Suiting local production and lifestyles
Adjustable housing and infrastructure
planning

No pre-disaster planning
Little supervision on the housing
quality
Few risk-coping mechanisms
Inadequate assessment due to rush of
progress
Inadequate infrastructure
Weak basis for future development

all news reports cover our research concerns. Various concerns have been paid in
the reconstruction news reports. This is naturally and logically as different journalists usually uses different views to investigate the post-disaster reconstruction. In
order to refine the useful information, the information of the first three high frequency was considered as key points in relevant concerns, as shown in the last
column of Table 1.
This study further analyzes the experiences and problems found in the mined
cases. Table 2 lists the top 10 experiences and problems respectively.

3.2 Discussions
The identified different problems and experiences were useful for experience
sharing. The most important experiences mainly lie in available matched policies,
successful organization, involving rural victims in decision making, and satisfying
rural victims’ needs. These issues are valuable considerations and should be considered in future reconstruction. The typical problems include no pre-disaster
planning, little supervision on the housing quality, insufficient financial support on
daily management for CRS, little attention to psychological recovery [18].
Diversified reasons result in the identified problems above. Typically, some
critical activities of developing CRS in post-disaster reconstruction are missing. For
example, although there were some emergency measures for responding to natural
disasters, no pre-disaster planning was available. As a result, no specific guidelines
can be provided to instruct post-disaster reconstruction. In addition, some activities
are not so well-designed due to the rush of the project. For example, there is little
supervision on the quality of housing and very little assessment on the reconstruction process. This may lead to critical issues if another disaster occurs [19–22].

4

Conclusion

This study adopts data mining to systematically investigate post-disaster reconstruction in rural China. Credible news websites were accessed to select the
reconstruction cases for analysis. Python was used to develop special programs to
mine the reconstruction cases. It was found that various concerns have been paid in
the reconstruction news reports. The most important experiences and problems have
also been identified for sharing. The results can also facilitate the government and
NGOs to find suitable measures to promote sustainable development in post-disaster
reconstruction. However, it should be noticed that detailed text analysis has not been
conducted due to limit of time. Typical text mining, such as text categorization and
text clustering should be conducted in future studies. In addition, the mining source
merely includes online news reports, which may also present some bias of the
results. Future studies may include the academic papers for mining.

Acknowledgement The work described in this paper was supported by the National Natural
Science Foundation of China (71503228).

References
1. EM-DAT (2013) The international disaster database. http://www.emdat.be/database (5 Jan 2013)
2. Zhang YZ (2010) The key issue of new socialist countryside construction is to develop rural
economy. J ShanXi Finance Econ Univ 32(2):160 (in Chinese)
3. Na BG, He ZY (2011) Enterprise-village’ rural reconstruction financing mode based on CSR
perspective. Rural Econ 11:61–64 (in Chinese)
4. Jin L (2008) Some urgent problems on reconstruction after the earthquake of Wenchuan.
Planners 7 (in Chinese)
5. Pelling M (2007) Learning from others: the scope and challenges for participatory disaster risk
assessment. Disasters 31(4):373–385
6. Cheng FF, Dong XM, Wang SY (2010) Emergency management of Yushu earthquake tests
the Wenchuan experience. J Evid Based Med 10(3):157–162 (in Chinese)
7. Quarantelli EL (1999) The disaster recovery process: what we know and do not know from
research. http://dspace.udel.edu:8080/dspace/bitstream/19716/309/1/PP%20286.pdf. Accessed
on 10 Aug 2010
8. Haas JE, Kates RW, Bowden MJ (1977) Reconstruction following disaster. MIT Press,
Cambridge
9. Rubin CB, Popkin R (1990) Disaster recovery after hurricane Hugo in South Carolina.
Institute of Behavioral Science, University of Colorado, Boulder
10. Brunsdon D, Smith S (2004) Summary notes from the infrastructure workshop. In: New
Zealand recovery symposium, The Ministry of Civil Defence and Emergency Management
11. Hutanuwatr K, Bolin B, Pijawka D (2012) Vulnerability and disaster in Thailand: scale,
power, and collaboration in post-tsunami recovery. In: Forces of nature and cultural responses.
Springer, Netherlands, pp 69–92
12. Nigg JM (1995) Disaster recovery as a social process. http://dspace.udel.edu:8080/dspace/
bitstream/19716/625/1/PP219.pdf. Accessed on 8 Aug 2010
13. Mileti D (1999) Disasters by design: a reassessment of natural hazards in the United States.
Joseph Henry Press, Washington DC
14. Miles SB, Chang SE (2006) Modeling community recovery from earthquakes. Earthquake
Spectra 22(2):439–458
15. Lyons M (2009) Building back better: the large-scale impact of small-scale approached to
reconstruction. World Dev 37(2):385–398
16. Diaz JOP, Dayal A (2008) Sense of place: a model for community based psychosocial support
programs. http://www.massey.ac.nz/*trauma/issues/2008-1/prewitt_diaz.htm. Accessed 11
Aug 2010
17. Rank of credible news websites. http://top.chinaz.com/list.aspx?t=247. Accessed on 15 June 2015
18. Berry MW (2004) Survey of text mining. Springer, Berlin
19. Peng Y (2015) A comparison of two approaches to develop concentrated rural settlements
after the 5.12 Sichuan Earthquake in China. Habitat Int 49:230–242
20. Peng Y, Shen QP, Shen LY, Lu C, Yuan Z (2014) A generic decision model for developing
concentrated rural settlement in post-disaster reconstruction: a China study. Nat Hazards 71
(1):611–637
21. Peng Y, Shen LY, Zhang XL, Ochoa JJ (2014) The feasibility of concentrated rural settlement
in a context of post-disaster reconstruction: a study of China. Disasters 38(1):108–124
22. Peng Y, Shen LY, Tan C, Tan DL, Wang H (2013) Critical determinant factors (CDFs) for
developing concentrated rural settlement in post disaster reconstruction: a China study. Nat
Hazards 66(2):355–373