Please wait a minute...
Shopping Cart Email List Chinese
Advanced Search Fig/Tab
Frontiers of Engineering Management

ISSN 2095-7513

ISSN 2096-0255(Online)

CN 10-1205/N

Postal Subscription Code 80-905

Featured Articles  |  Most Downloaded  |  Most Reading
Online Submission Subscribe to Our RSS Content Feed
Front. Eng    2019, Vol. 6 Issue (2) : 262-274    https://doi.org/10.1007/s42524-019-0015-6
RESEARCH ARTICLE
Analyzing construction safety through time series methods
Houchen CAO1(), Yang Miang GOH2
1. Department of Building, School of Design and Environment, National University of Singapore, 4 Architecture Dr., 117566, Singapore
2. Safety and Resilience Research Unit (SaRRU), Department of Building, School of Design and Environment, National University of Singapore, 4 Architecture Dr., 117566, Singapore
 Download: PDF(613 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

The construction industry produces a large amount of data on a daily basis. However, existing data sets have not been fully exploited in analyzing the safety factors of construction projects. Thus, this work describes how temporal analysis techniques can be applied to improve the safety management of construction data. Various time series (TS) methods were adopted for identifying the leading indicators or predictors of construction accidents. The data set used herein was obtained from a large construction company that is based in Singapore and contains safety inspection scores, accident cases, and project-related data collected from 2008 to 2015. Five projects with complete and sufficient data for temporal analysis were selected from the data set. The filtered data set contained 23 potential leading indicators (predictors or input variables) of accidents (output or dependent variable). TS analyses were used to identify suitable accident predictors for each of the five projects. Subsequently, the selected input variables were used to develop three different TS models for predicting accident occurrences, and the vector error correction model was found to be the best model. It had the lowest root mean squared error value for three of the five projects analyzed. This study provides insights into how construction companies can utilize TS data analysis to identify projects with high risk of accidents.
Keywords time series      temporal      construction safety      leading indicators      accident prevention      forecasting     
Corresponding Author(s): Houchen CAO   
Online First Date: 04 March 2019    Issue Date: 17 May 2019
 Cite this article:   
Houchen CAO,Yang Miang GOH. Analyzing construction safety through time series methods[J]. Front. Eng, 2019, 6(2): 262-274.
 URL:  
https://academic.hep.com.cn/fem/EN/10.1007/s42524-019-0015-6
https://academic.hep.com.cn/fem/EN/Y2019/V6/I2/262
Fig.1  Time series analysis flowchart
Models Description Requirements Type
Moving average
(MA)		 A method used for smoothing TS; usually a series of arithmetic means (Wohlrabe and Mittnik, 2016). TS needs to be stationary. Linear model
Exponential smoothing (ES) A form of weighted MA whereby the weights decrease exponentially over time (Wohlrabe and Mittnik, 2016). Famous models include Holt–Winters (HW). TS can be nonstationary. Can be linear or nonlinear
Autoregression
(AR)		 Uses correlations of successive lagged observations of itself to forecast values (Wohlrabe and Mittnik, 2016). TS can be nonstationary, thus resulting in a random walk model. Linear model
ARMA Combines AR and MA to generate an accurate predictive stochastic model. TS needs to be stationary. Linear model
Box–Jenkins AR Integrated MA
(ARIMA)		 ARIMA allows for the inclusion of nonstationary TS data in automatic modeling. Here, the “I” of the model stands for integration: the reverse process of differencing data to achieve stationarity (Brockwell and Davis, 2002).. TS can be nonstationary. Linear model
Tab.1  Common univariate TS models
Fig.2  Multivariate TS model selection flowchart
Topic Univariate Models Multivariate Models Others
Safety • VAR Model (Lingard et al., 2017) • Complex Network Theory (Zhou et al., 2017)
Construction Productivity • ARIMA Model (Kim et al., 2015)
Project Progress • ES (Batselier and Vanhoucke, 2017) • Grey Forecasting Model (Lin et al., 2012)
Manpower • Linear Regression (Bell and Brandenburg, 2003) • VEC Model (Wong et al., 2011)
• VEC Model (Wong et al., 2007)
• Multiple Linear Regression (Wong et al., 2008)		 • Grey Forecasting Model (Ho, 2010)
Construction Demand • Box–Jenkins (ARIMA) Model (Fan et al., 2010) • VAR Model (Sing et al., 2015) • Grey Forecasting Model (Tan et al., 2015)
• Combination of SVM, NNM, and ARIMA models (Lam and Oshodi, 2016)
Construction Costs • Regression integrated with ARIMA (Ng et al., 2004)
• ARIMA Model (Hwang et al., 2012)
• Comparative Study between Simple MA (SMA), ARIMA, and HW ES (Ashuri and Lu, 2010)		 • VEC Model (Wong and Ng, 2010)
• VEC Model (Shahandashti and Ashuri, 2013)
• VEC Model (Shahandashti and Ashuri, 2016)		 • Data Clustering and Pooling (Yeung and Skitmore, 2012)
• Neural Networks (Cao et al., 2015)
• Comparative Study using Multiple TS Models (ARMA, VAR, Neural Networks, MA, HW ES) (Hwang, 2011)
Tab.2  Various TS methods used in current literature
Fig.3  Research design
Records Description Potential Indicators to be Selected Attribute Type
Construction Incident Summary This set of records consists of the number of accidents that occurred each month, with the exact date and nature of injuries. Number of accidents per month Ratio
Monthly Project Manager (PM) Inspection Checklist The monthly inspection checklist is crucial for this research. Each PM inspection checklist contains a set of standard prescribed elements, which contribute directly to the safety of the project (e.g., PPE, overhead protection, falling hazards, and hazardous materials).
During each monthly inspection, the PM inspection team gives a score to each element, and a final weighted score that totals the scoring of all elements is generated.		 Every element found in the inspection checklist Interval
Monthly Project Progress This data set is a record of each project’s progress and delay per month for Company X. Project monthly progress and delay Ratio
Monthly Project Manpower Statistics This data set comprises the exact number of man hours on site per month for each project under Company X. Project man hours per month Ratio
Tab.3  Description of data files
Project Data set Acronym Project Type Project Ownership Contract Sum (S$) Sets of Observations Available
(in Months)
CVL01 Civil works
(Train rails)		 Public 177.58 M 50
CVL02 Civil works
(Train rails)		 Public 268.68 M 48
BDG01 Building
(Redevelopment)		 Private 101.3 M 28
BDG02 Building
(Redevelopment)		 Private 204.8 M 48
A&A01 Building
(Additions & Alterations)		 Private 110.5 M 31
Tab.4  Description of projects selected for modeling
Number Potential Leading Indicators Description
1 Number of Accidents (Accidents) Number of accidents in each project per month
2 Project Progress (Prog) Current project progress, in percentage
3 Project Delay (Delay) Difference between current and planned progress
4 Project Man Hours (ManHours) Amount of man hours input for a project
5 Overall PM Inspection Score (PMScore) Weighted sum of all inspection elements
6 Inspection Element: Personal Protective Equipment (PPE) Wearing of protective equipment (safety harness, belt, lifeline, ear plug, safety boots, face mask, etc.)
7 Inspection Element: Overhead Protection (OHPro) Provision of periphery shelter and exit/entry points; area exposed to falling objects
8 Inspection Element: Excavation Work (ExWork) Provision of shoring systems, warning signage, and barricades along perimeter of excavation
9 Inspection Element: Machine Safety Guarding (MSG) Provision of safety guards at moving parts, blades, and similar parts of machinery
10 Inspection Element: Safe Means of Access (SafeAccess) Access that is free from obstructions and tripping hazards; provision of lighting and directional signage
11 Inspection Element: Operating Crane / Lifting (Crane) Presence of supervisor, signalman, and rigger assist; identification of lifting crew, etc.
12 Inspection Element: Scaffold (Scaffold) Provision of standard erection members, scaffolding access, and load signboard
13 Inspection Element: Tower / Mobile Scaffold (MobileScaf) Similar to above: Scaffold
14 Inspection Element: Mech Elevated Work Platform (MEWP) Provision of working permits; presence of warning signs
15 Inspection Element: Falling Hazard / Opening (FallHaz) Provision of barricades to potential openings on site
16 Inspection Element: Electrical Hazard (ElecHaz) Proper installation of cables; provision of suitable insulation
17 Inspection Element: First-aid Facilities (FirstAid) Provision of first-aid box and room on site
18 Inspection Element: Emergency Preparedness (EP) Provision of firefighting equipment; conduct of emergency drills with maintained records
19 Inspection Element: Handling & Storage of Hazardous Materials (HazMat) Proper disposal of hazardous waste; provision of instructions and manuals
20 Inspection Element: Safe Work Procedure (SWP) Proper procedures for working at height, confined spaces, hot works, etc.
21 Inspection Element: Power Tool Safety (PTS) Proper and safe use of electric tools
22 Inspection Element: Earth Control Measures (ECM) Proper procedures for preventing silt pollution of surrounding environment
23 Inspection Element: Noise, Vector, and Others (NVO) Proper procedures that minimize or prevent excessive noise, vector, etc.
Tab.5  Finalized pool of potential leading indicators in each set of project data
Fig.4  Summary of Stage 2 data processing
Indicators Coefficient “r” Test Statistic P-value
Accidents, Prog 0.439 3.3864 0.001421*
Accidents, Delay 0.444 3.4302 0.001249*
Accidents, ManHours 0.356 2.6418 0.0111*
Accidents, PMScore −0.196 −1.3816 0.1735
Accidents, PPE −0.00985 −0.06827 0.9458
Accidents, OHPro 0.0910 0.63028 0.5315
Accidents, ExWork 0.147 1.0285 0.3089
Accidents, MSG 0.221 1.573 0.1223
Accidents, SafeAccess −0.095 −0.664 0.5101
Accidents, Crane 0.0440 0.3055 0.7613
Accidents, Scaffold −0.352 −2.6095 0.01206*
Accidents, MobileScaf −0.332 −2.4387 0.01849*
Accidents, MEWP −0.0715 −0.49669 0.6217
Accidents, FallHaz −0.06929 −0.48121 0.6326
Accidents, ElecHaz −0.2308 −1.6439 0.1067
Accidents, FirstAid 0.2016 1.4264 0.1602
Accidents, EP −0.06564 −0.45578 0.6506
Accidents, HazMat −0.01506 −0.1044 0.9173
Accidents, SWP 0.1242 0.86748 0.39
Accidents, PTS 0.1724 1.2127 0.2312
Accidents, ECM 0.098 0.68226 0.4984
Accidents, NVO 0.10199 0.71034 0.4809
Tab.6  CVL01 correlation tests between “accidents” and other indicators
Indicators Tested with D Accidents Lag Length
2 4 6
D Accidents &D PMScore 0.2515 0.6246 0.9027
D Accidents &D PPE 0.4134 0.8031 0.6931
D Accidents &D OHPro 0.003189* 0.01228* 0.02892*
D Accidents &D ExWork 0.3697 0.5058 0.8327
D Accidents &D MSG 0.6334 0.7974 0.7699
D Accidents &D SafeAccess 0.8179 0.004176* 0.02366*
D Accidents &D Crane 0.7892 0.1742 0.2549
D Accidents &D Scaffold 0.09519 0.1863 0.1918
D Accidents &D MobileScaf 0.129 0.2786 0.3819
D Accidents &D MEWP 0.9037 0.8585 0.6584
D Accidents &D FallHaz 0.1371 0.3161 0.6377
D Accidents &D ElecHaz 0.2972 0.1994 0.634
D Accidents &D FirstAid 0.3255 0.4529 0.2639
D Accidents &D EP 0.1301 0.1069 0.2292
D Accidents &D HazMat 0.14 0.0748 0.005423*
D Accidents &D SWP 0.1307 0.1619 0.1267
D Accidents &D PTS 0.04684 0.2966 0.3721
D Accidents &D ECM 0.9679 0.8669 0.6728
D Accidents &D NVO 0.2237 0.4701 0.7294
Tab.7  CVL01 Granger causality test against accidents
Test A: Results for Accidents and OHPro, SafeAccess, HazMat
Number of Cointegrating Relationships r Trace Statistic 0.05 Critical Value Probability
None 65.77 47.86 0.0005*
At most 1 37.01 29.80 0.0062*
At most 2 21.73 15.49 0.0050*
At most 3 9.52 3.84 0.0020*
Tab.8  CVL01 Cointegration test results of various indicators
Test B: Results for Accidents and OHPro
Number of Cointegrating Relationships r Trace Statistic 0.05 Critical Value Probability
None 22.30 15.50 0.0040*
At most 1 8.06 3.84 0.0045*
Tab.9  
Test C: Results for Accidents and SafeAccess
Number of Cointegrating Relationships r Trace Statistic 0.05 Critical Value Probability
None 17.54 15.49 0.0243*
At most 1 4.80 3.84 0.0284*
Tab.10  
Test D: Results for Accidents and HazMat
Number of Cointegrating Relationships r Trace Statistic 0.05 Critical Value Probability
None 16.74 15.49 0.0323*
At most 1 4.96 3.84 0.0260*
Tab.11  
Statistics Values
R-squared Value 0.70
Forecast RMSE 1.583
Tab.12  Regression and forecast values for VEC model in cvl01
Statistics Values
Final Model ARIMA (0, 1, 1)
ARIMA Forecast RMSE 1.587
Tab.13  ARIMA forecast statistics of cvl01
Statistics Values
Hidden Layers 2 layers, 10 nodes – 2 nodes
Forecast RMSE 0.6858
Tab.14  ANN forecast statistics of cvl01
Fig.5  ANN representation
Project Selected Input Variables
CVL01 (Train rails – Civil) Accidents, OHPro, SafeAccess, HazMat
CVL02 (Train rails – Civil) Accidents, PMScore, MobileScaf, EP
BDG01 (Redevelopment – Building) Accidents, Crane, MobileScaf, SWP
BDG02 (Redevelopment – Building) Accidents, PMScore, MobileScaf, EP
A&A01 (A&A – Building) Accidents, SafeAccess
Tab.15  Leading indicators of each project
Project VEC Model RMSE
/ R-squared Value		 ARIMA
RMSE		 ANN
RMSE
CVL01
(Train rails –Civil)		 1.583/0.70 1.587 0.6858*
CVL02
(Train rails –Civil)		 1.355/0.66* 1.717 1.637
BDG01 (Redevelopment –Building) 1.428/0.75 0.9110* 1.483
BDG02 (Redevelopment – Building) 0.9785/0.66* 1.717 1.261
A&A01
(A&A – Building)		 0.4710/0.74* 0.9724 1.862
Tab.16  Forecasting model statistics for each project
1 N KAhmed, A F Atiya, N E Gayar, H El-Shishiny (2010). An empirical comparison of machine learning models for time series forecasting. Econometric Reviews, 29(5–6): 594–621
https://doi.org/10.1080/07474938.2010.481556
2 BAshuri, J Lu (2010). Time series analysis of ENR construction cost index. Journal of Construction Engineering and Management, 136(11): 1227–1237
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000231
3 JBatselier, M Vanhoucke (2017). Improving project forecast accuracy by integrating earned value management with exponential smoothing and reference class forecasting. International Journal of Project Management, 35(1): 28–43
https://doi.org/10.1016/j.ijproman.2016.10.003
4 L CBell, S G Brandenburg (2003). Forecasting construction staffing for transportation agencies. Journal of Management Engineering, 19(3): 116–120
https://doi.org/10.1061/(ASCE)0742-597X(2003)19:3(116)
5 G E PBox, G MJenkins (1976). Time Series Analysis Forecasting and Control. California: Prentice-Hall International, Inc.
6 P JBrockwell, R ADavis (2002). Introduction to Time Series Forecasting. New York: Springer
7 A JBuck (1999). Cointegration and error correction.
8 MCao, M Cheng, YWu (2015). Hybrid computational model for forecasting taiwan construction cost index. Journal of Construction Engineering and Management, 141(4): 04014089
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000948
9 L CChao, M J Skibniewski (1995). Neural network method of estimating construction technology acceptability. Journal of Construction Engineering and Management, 121(1): 130–142
https://doi.org/10.1061/(ASCE)0733-9364(1995)121:1(130)
10 R Y CFan, S TNg, J M WWong (2010). Reliability of the Box–Jenkins model for forecasting construction demand covering times of economic austerity. Construction Management and Economics, 28(3): 241–254
https://doi.org/10.1080/01446190903369899
11 C W JGranger (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3): 424–438
https://doi.org/10.2307/1912791
12 P H KHo (2010). Forecasting construction manpower demand by gray model. Journal of Construction Engineering and Management, 136(12): 1299–1305
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000238
13 SHwang (2011). Time series models for forecasting construction costs using time series indexes. Journal of Construction Engineering and Management, 137(9): 656–662
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000350
14 SHwang, M Park, HLee, HKim (2012). Automated time-series cost forecasting system for construction materials. Journal of Construction Engineering and Management, 138(11): 1259–1269
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000536
15 R JHyndman, G Athanasopoulos (2013). Forecasting: Principles and Practice.
42 Global Inc.HS (2017). EViews 9.5 Student Version Lite. California, USA: IHS Inc.
23 OrganizationInternational Labour (2018). The enormous burden of poor working conditions.
16 HKim, H S Lee, M Park, C RAhn, SHwang (2015). Productivity forecasting of newly added workers based on time-series analysis and site learning. Journal of Construction Engineering and Management, 141(9): 05015008
https://doi.org/10.1061/(ASCE)CO.1943-7862.0001002
17 K CLam, O S Oshodi (2016). Using univariate models for construction output forecasting: Comparing artificial intelligence and econometric techniques. Journal of Management Engineering, 32(6): 04016021
https://doi.org/10.1061/(ASCE)ME.1943-5479.0000462
18 C MLin, P H Tserng, P S Ho, D L Young (2012). A novel dynamic progress forecasting approach for construction projects. Expert Systems with Applications, 39(3): 2247–2255
https://doi.org/10.1016/j.eswa.2011.07.093
19 HLingard, M Hallowell, RSalas, PPirzadeh (2017). Leading or lagging? Temporal analysis of safety indicators on a large infrastructure construction project. Safety Science, 91: 206–220
https://doi.org/10.1016/j.ssci.2016.08.020
20 ALund, M Lund (2013). Pearson’s correlation using stata.
21 RNau (2017). Statistical forecasting: Notes on regression and time series analysis.
22 S TNg, S O Cheung, M Skitmore, T C YWong (2004). An integrated regression analysis and time series model for construction tender price index forecasting. Construction Management and Economics, 22(5): 483–493
https://doi.org/10.1080/0144619042000202799
24 BPfaff (2008). Analysis of Integrated and Cointegrated Time Series with R. 2nd ed. New York: Springer
25 C Q XPoh, C UUbeynarayana, Y MGoh (2018). Developing safety leading indicators for construction sites: A machine learning approach. Automation in Construction
26 B H KSchabowicz, BHoła (2008). Application of artificial neural networks in predicting earthmoving machinery effectiveness ratios. Archives of Civil and Mechanical Engineering, 8(4): 73–84
https://doi.org/10.1016/S1644-9665(12)60123-X
27 S MShahandashti, BAshuri (2013). Forecasting engineering news-record construction cost index using multivariate time series models. Journal of Construction Engineering and Management, 139(9): 1237–1243
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000689
28 S MShahandashti, BAshuri (2016). Highway construction cost forecasting using vector error correction models. Journal of Management Engineering, 32(2): 04015040
https://doi.org/10.1061/(ASCE)ME.1943-5479.0000404
29 M C PSing, D JEdwards, H J XLiu, P E DLove (2015). Forecasting private-sector construction works: VAR model using economic indicators. Journal of Construction Engineering and Management, 141(11): 04015037
https://doi.org/10.1061/(ASCE)CO.1943-7862.0001016
30 YTan, C Langston, MWu, J JOchoa (2015). Grey forecasting of construction demand in Hong Kong over the next ten years. International Journal of Construction Management, 15(3): 219–228
https://doi.org/10.1080/15623599.2015.1066570
31 The R Development Core Team (2018). R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing
32 SUmit Dikmen, MSonmez (2010). An artificial neural networks model for the estimation of formwork labour. Journal of Civil Engineering and Management, 17(3): 340–347
https://doi.org/10.3846/13923730.2011.594154
33 KWohlrabe, S Mittnik (2016). Univariate time series analysis.
35 J M WWong, A P CChan, Y HChiang (2007). Forecasting construction manpower demand: A vector error correction model. Building and Environment, 42(8): 3030–3041
https://doi.org/10.1016/j.buildenv.2006.07.024
34 J M WWong, A P CChan, Y HChiang (2008). Modeling and forecasting construction labor demand multivariate analysis. Journal of Construction Engineering and Management, 134(9): 664–672
https://doi.org/10.1061/(ASCE)0733-9364(2008)134:9(664)
36 J M WWong, A P CChan, Y HChiang (2011). Construction manpower demand forecasting. Engineering, Construction, and Architectural Management, 18(1): 7–29
https://doi.org/10.1108/09699981111098667
37 J M WWong, S TNg (2010). Forecasting construction tender price index in Hong Kong using vector error correction model. Construction Management and Economics, 28(12): 1255–1268
https://doi.org/10.1080/01446193.2010.487536
38 DYeung, M Skitmore (2012). A method for systematically pooling data in very early stage construction price forecasting. Construction Management and Economics, 30(11): 929–939
https://doi.org/10.1080/01446193.2012.733402
39 G QZhang, B E Patuwo, M Y Hu (1997). Forecasting with artificial neural networks: The state of the art. International Journal of Forecasting, 14: 35–62
https://doi.org/10.1016/S0169-2070(97)00044-7
40 CZhou, L Ding, M JSkibniewski, HLuo, S Jiang (2017). Characterizing time series of near-miss accidents in metro construction via complex network theory. Safety Science, 98: 145–158
https://doi.org/10.1016/j.ssci.2017.06.012
41 ZZhou, Y M Goh, Q Li (2015). Overview and analysis of safety management studies in the construction industry. Safety Science, 72: 337–350
https://doi.org/10.1016/j.ssci.2014.10.006
[1] Benachir MEDJDOUB, Moulay Larbi CHALAL. Impact of household transitions on domestic energy consumption and its applicability to urban energy planning[J]. Front. Eng, 2017, 4(2): 171-183.
[2] Fei Dai,Yoojung Yoon,Hota V. GangaRao. State of Practice of Construction Site Safety in the USA[J]. Front. Eng, 2016, 3(3): 275-282.
[3] Xiao-hong Chen,Xiang-bo Tang. Spatial and Temporal Distribution of PM2.5 Pollution and Control Strategy in the Chang-Zhu-Tan City Cluster[J]. Front. Eng, 2016, 3(2): 171-181.
[4] Shu-lin Zhang,Hui-qing Jin,Yang Song,Wan-sheng Yu,Liang-dan Sun. Genetic Study Identifies CBLN4 as a Novel Susceptibility Gene for Accident Proneness[J]. Front. Eng, 2016, 3(1): 30-38.
[5] Curt Webb,Lu Gao,Ling-guang Song. Schedule Compression Impact on Construction Project Safety[J]. Front. Eng, 2015, 2(4): 344-350.
[6] Lin Yang,Qi-ming Li. The Research Review of the Impact of Design to Construction Safety[J]. Front. Eng, 2015, 2(3): 224-227.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed