© 2013 Pakistan Journal of Statistics 611 Pak. J. Statist.
2013 Vol. 29(5), 611-620
USING DATA MINING TECHNOLOGY TO EXPLORE LABOR SAFETY STRATEGY- A LESSON FROM THE CONSTRUCTION INDUSTRY
Sung-Lin Hsueh1, Chung-Fah Huang2§ and Cheng-Yung Tseng2
1 Graduate Institute of Cultural and Creative Design,
Tung Fang Design Institute, No.110 Dongfang Road, Hunei Distract, Kaohsiung City 82941, Taiwan.
2 Dept of Civil Engineering, National Kaohsiung University
of Applied Sciences, No 415, Chien-Kung Rd., Kaohsiung City, Taiwan.
§
Corresponding author. Email: jeffrey@kuas.edu.tw ABSTRACT
The working environments of the construction industry in Taiwan generally have higher risks, accounting for over 50% of major occupational injuries in Taiwan. In this study, data mining was used to analyze totally 33,298 cases of labor safety inspection in labor safety of Taiwan’s construction industry. This research found insufficient fall prevention; construction safety management and electric shock prevention will lead to higher likelihood of construction safety hazards. In addition, the incorporation of Data Mining technology in the labor safety inspection system will help to reduce the likelihood of construction hazards and improve the effectiveness and results of safety management.
KEY WORDS
Construction Industry; Data Mining; Labor Safety Inspection; Safety Management; Disaster Management.
1. INTRODUCTION
The construction industry is known for its massive output value. However, according to Moodley et al. (2008), it is also an industry with low moral reputation with rampant problems of corruption and bribery. In addition, it can pose threat to the health and safety of not only humans but also the environment if the construction activities are not properly managed and monitored. Furthermore, there is a great risk and cost of accidents in the construction industry (Larsson et al., 2002, Moon, 2007). It indicates in ILO (2003), about 60,000 people die on construction projects. According to Bureau of Labor Statistics (2005, 2009) and MOM (2007) accident statistics of construction workers shows as Table 1. The total workplace fatality rates (per 100 thousand workers) are shown in Table 2. The fatality rates in the construction sector (per 100 thousand workers).
Based on the estimation by the United Nations Environment Programmer, the construction industry was about using the 40% of global energy (Lam et al., 2010). The building sector has been accused of causing the pollution of the environment (Ding, 2008). In addition, the construction industry has the structural characteristic of temporary organizations (Koskela, 2003) while different environmental factors, competition as
Using Data Mining Technology To Explore Labor Safety strategy… 612
severe as the competition in the retail industry (Cheng et al., 2001) and multiple levels of outsourcing all together make it more difficult to ensure good for management of construction industry. This explains the high frequency of vocational injuries in the construction industry.
It is the laws of occupational safety and health in Taiwan, The safety of labor is the duty of the employer. Over the past years, in the major causalities caused by vocational hazards in Taiwan, there are over 45%
related to the construction industry (Council of Labor Affairs, 2012). Many Research on Taiwan’s construction projects. It indicates the gravity of occupational injuries (Chi et al., 2004; Liao and Perng, 2008; Lin et al., 2008; and Cheng, 2012) and had 1,230 occupational injury that 50.6% accidents occurred in the construction projects (Chi and Wu, 1997, Cheng, 2010). From the statistics for person-time in the vocational hazards in the construction industry (Council of Labor Affairs, 2012), it indicates in Taiwanese construction the person-time in vocational hazards has no significant decrease in from 2001 to 2011 (as Table 3). According to this high figure indication, the construction industry in Taiwan requires more effective construction safety management.
To reduce labor safety problems in the construction industry, the Regulatory and supervisory authorities mainly make inspection in accordance with governing laws and regulations. Those legal regulations include such as the Labor Inspection Act, Labor Standards Act, Labor Safety and Health Act and several others. These laws and regulations that cover a wide variety of issues and labor inspections could somehow help to improve the facilities and organizations in charge of construction safety management. From a massive amount of inspection data was explored and analyzed, there is still no effective method to measure the inspections have a direct influence on reducing vocational hazards in the Taiwanese construction industry. The factors of inherently hazardous nature, environment and equipment, personnel and management are the major causes of high accident rates in construction industry (Tam et al., 2004; Fabiano et al., 2004).
Table 3: The Vocational Hazards in the Construction Industry Person-Time, from 2001 to 2011, in Taiwan
Year Total Disease & Injury Disability Death Notes
2001 11,115 10,062 868 185 1. Data resource: Bureau of Labour Insurance. 2. The traffic accidents are included. 2002 10,439 9,447 827 165 2003 10,304 9,462 671 171 2004 10,783 10,003 658 122 2005 10,528 9,694 669 165 2006 11,222 10,401 654 167 2007 11,327 10,516 687 124 2008 11,823 11,113 594 116 2009 11,588 10,890 580 118 2010 11,660 11,001 553 106 2011 11,756 11,045 590 121
Table 1: The Total Workplace Fatality Rates (Per 100 Thousand Workers)
2005 2006 2007 United States 4% 4% 3.8%
Singapore 4% 3.1% 2.9%
Table 2: The Fatality Rates in the Construction Sector (Per 100 Thousand Workers)
2005 2006 2007 United States 11.1% 10.9% 10.5% Singapore 11.9% 9.4% 8.1%
Hsueh, Huangand Tseng 613 Beside the above background and motivation, in the past the researches in accidents in labour safety and accidents common focus on specified constructing conditions or analyses in accidents. These are mostly based on Qualitative Research (Case study), reviewing results (from accidents) and reasons and factors, but only few in statistics quantitative researches. The major focuses in those statistics quantitative researches are the conditions (pressure) of the construction labours, but never on labours’ safety inspection. From above reasons, the major aims of this research are: (1) To analyse the common deficiencies and defects in labours’ safeties, and understand the situations of deficiencies and defects in current. (2) Analyzing the relations between accidents and (Labours’ safeties relative) legal offense, to understand the possible negligence in regulations by official labours inspectors, for inspected companies to have better understandings between safety issues and the matters need attention. (3) Reviewing official inspections how to influence construction safety accidents, and analyse the efficacy in inspections; from established model to predict potential accidents and then find out and give more supervision the construction projects have high risk in labours` safety, and more requirements in safety improvements.
Data mining is a great method to analyze massive amounts of data. It has been widely applied in information systems as an important tool to automatically analyze data and turn data into knowledge. However, the application of data mining in the construction industry is rarely explored and discussed (Lee et al., 2008). Data mining enables the decision maker to find out information patterns hidden inside massive and seemly unrelated data and extract reference information for his decision making through the graphic expressions of the connection-based and rule-based data mining results (Huang and Hsueh, 2010).
The past studies on labor safety incidents mostly focus on analyzing specific construction projects or incidents. Among the few studies that are based on quantitative and statistical methods, they all focus on analysis of the physical and psychological conditions (such as stress) of the construction workers without any discussion about labor safety inspections. Therefore, this research is a preliminary attempt to use data mining to discuss labor safety inspections and their effect on reducing vocational hazards of the construction industry.
2. METHODOLOGY AND DATA PROCESSING
In this study, the data of labor inspections on the large construction projects in southern Taiwan over the past two years were collected and analyzed using data mining. In addition, the data mining results were further discussed with senior labor safety inspectors.
2.1 Research Data
We used the data of inspection results of the Construction Section (Labor Inspection Office of Southern Taiwan). There were totally 33,298 entries of inspection data, some of which were inspections of the same large construction projects. There are many items of information in each entry. In this research, only the information items of inspection date, construction project title, violated laws and articles, and treatment results were used. Table 4 is an extraction of the raw data used in this study.
Hsueh, Huangand Tseng 619 4. CONCLUSION
According to the findings of this research, failures to meet the legal requirements about fall, electric shock and collapse preventions are the problems commonly in the construction projects over the past years. Such failures are also major reason for casualties in vocation hazards. Therefore, the governing authorizes should focus more on these problems. There are three scenarios that are likely to cause serious consequences. The first scenario covers accidents caused by inadequate fall prevention. The second scenario covers accidents caused by inadequate safety management by the constructor despite the existence of adequate fall prevention. The third scenario covers accidents caused by inadequate electric shock prevention. Therefore, it is suggested the governing authorities to strictly require the installment and use of fall prevention facilities and gears at the construction sites to reduce injuries and casualties.
This research is an unprecedented attempt to use data mining technology to analyze the connections between labor safety inspections and related legal regulations with the goal of establishing a model to predict possible construction safety hazards. If the model can be incorporated into the existing labor safety inspection information system, it will help to better analyze the data. In addition, the model can be used in the inspections on the projects still under construction to efficiently prevent possible hazards by finding patterns similar to those had caused hazards in the past.
REFERENCES
1. Moodley, K., Smith, N.J. and Preece, C.N. (2008). Stakeholder matrix for ethical relationships in the construction industry. Const. Mgmt. and Econ., 26(6), 625-632. 2. Larsson, T.J. and Field, B. (2002). The distribution of occupational injuries risks in
the Victorian construction industry. Safety Science, 40, 439-456.
3. Moon, S.W., Kim, J.S. and Kwon, K.N. (2007). Effectiveness of OLAP-based cost data management in construction cost estimate. Automation in Construction, 16, 336-344.
4. International Labour Organization (ILO), (2003). Safety in Numbers: Pointers for a
Global Safety at Work. International Labour Office, Geneva, Switzerland.
5. Bureau of Labor Statistics (2005). Census of Fatal Occupational Injuries 1990-2005. Department of Labor, Washington D.C.
6. Bureau of Labor Statistics (2009). Census of Fatal Occupational Injuries 1992-2007. Department of Labor, Washington D.C.
7. Ministry of Manpower, MOM (2007). Information and Statistics. Occupational Safety and Health Division: 2007 Report.
8. Lam, T.I., Chan, H.W., Chau, C.K., Poon, C.S. and Chun, K.P. (2010). Factors affecting the implementation of green specifications in construction. J. Environ.
Mgmt., 91(3), 654-661.
9. Ding, G.K.C. (2008). Sustainable construction- The role of environmental assessment tools. J. Environ. Mgmt., 86(3), 451-464.
10. Koskela, L. (2003). Is structural change the primary solution to the problems of construction? Build. Res. & Infor., 31(2), 85-96.
11. Cheng, E.W.L., Li, H., Love, P.E.D. and Irani, Z. (2001). An e-business model to support supply chain activities in construction. Logistics Infor. Mgmt., 14(1/2), 68-77.
Using Data Mining Technology To Explore Labor Safety strategy… 620
12. Council of Labor Affairs (2012). Executive Yuan, R.O.C. http://www.cla.gov.tw/ 13. Chi, C.F., Chang, T.C. and Hung, K.H. (2004). Significant industry-source of
injury-accident type for occupational fatalities in Taiwan. Int.l of Indus. Ergo., 34, 77-91. 14. Liao, C.W. and Perng, Y.H. (2008). Data mining for occupational injuries in Taiwan
construction industry. Safety Science, 46, 1091-1102.
15. Lin, Y.H., Chen, C.Y. and Luo, J.L. (2008). Gender and age distribution of occupational fatalities in Taiwan. Accident Analysis and Prevention, 40, 1604-1610. 16. Cheng, C.W., Leu, S.S., Cheng, Y.M., Wu, T.C. and Lin, C.C.(2012).Applying data
mining techniques to explore factors contributing to occupational injuries in Taiwan's construction industry. Accident Analysis & Prevention, 48, 214-222.
17. Chi, C.F. and Wu, M.L. (1997). Fatal occupational injuries in Taiwan - relationship between fatality rate and age. Safety Science, 27, 1-17.
18. Cheng, C.W., Lin, C.C. and Leu, S.S. (2010). Use of association rules to explore cause-effect relationships in occupational accidents in the Taiwan construction industry, Safety Science, 48(4), 436-444.
19. Tam, C.M., Zeng, S.X. and Deng, Z.M. (2004). Identifying elements of poor construction safety management in China. Safety Science, 42, 569-586.
20. Fabiano, B., Currò, F. and Pastorino, R. (2004). A study of the relationship between occupational injuries and firm size and type in the Italian industry. Safety Science, 42, 87-600.
21. Lee, J.R., Hsueh, S.L. and Tseng, H.P. (2008). Utilizing Data Mining to Discover Knowledge in Construction Enterprise Performance Records. Journal of Civil
Engineering Management, 14(2), 79-84.
22. Huang, C.F. and Hsueh, S.L. (2010). Customer Behavior and Decision Making in the Refurbishment Industry-A Data Mining Approach. Journal of Civil Engineering
Management, 16(1), 75-84.
23. Fayyad, U., Shapiro G.P. and Smyth P. (1996). The KDD process for extracting useful knowledge from volumes of data. Communication of the ACM, 39(11), 27-34. 24. Han, J. and Kamber, M. (2001). Data Mining: Concept and Techniques, Morgan
Kaufmann Publishers, San Francisco.
25. Chang, L.Y. and Chen, W.C. (2005). Data mining of tree-based models to analyze freeway accident frequency. J. Safety Res., 36, 365-375.
