1

Big Data in the Construction Industry: A Review of

Present Status, Opportunities, and Future Trends
1 Abstract-The ability to process large amounts of data and computing), compressed, in diverse proprietary formats, and 51

2 to extract useful insights from data has revolutionised society. intertwined [6]. Accordingly, this diverse data is collated in 52
3 This phenomenon-dubbed as Big Data-has applications for a federated BIM models, which are enriched gradually and 53
wide assortment of industries, including the construction industry.
persisted beyond the end-of-life of facilities. BIM files can
4
54
5 The construction industry already deals with large volumes of
6 heterogeneous data; which is expected to increase exponentially quickly get voluminous, with the design data of a 3-story 55

7 as technologies such as sensor networks and the Internet of building model easily reaching 50GB in size [7]. Noticeably, 56
8 Things are commoditized. In this paper, we present a detailed this data in any form and shape has intrinsic value to the 57
9 survey of the literature, investigating the application of Big Data performance of the industry. With the advent of embedded 58
techniques in the construction industry. We reviewed related
devices and sensors, facilities have even started to generate
10
59
11 works published in the databases of American Association of
12 Civil Engineers (ASCE), Institute of Electrical and Electronics massive data during the operations and maintenance stage, 60

13 Engineers (IEEE), Association of Computing Machinery (ACM), eventually leading to more rich sources of Big BIM Data. This 61
14 and Elsevier Science Direct Digital Library. While the application vast accumulation of BIM data has pushed the construction 62
of data analytics in the construction industry is not new, the
15
16 adoption of Big Data technologies in this industry remains at a industry to enter the Big Data era. 63

17 nascent stage and lags the broad uptake of these technologies in Big Data has three defining attributes (a.k.a. 3V's), namely 64

18 other fields. To the best of our knowledge, there is currently no (i) volume (terabytes, petabytes of data and beyond); (ii) 65
19 comprehensive survey of Big Data techniques in the context of variety (heterogeneous formats like text, sensors, audio, video, 66
20 the construction industry. This paper fills the void and presents a graphs and more); and (iii) velocity (continuous streams of the 67
wide-ranging interdisciplinary review of literature of fields such
21
22 as statistics, data mining and warehousing, machine learning, and data). The 3V's of Big Data are clearly evident in construction 68

23 Big Data Analytics in the context of the construction industry. data. Construction data is typically large, heterogeneous, and 69

24 We discuss the current state of adoption of Big Data in the dynamic [8]. Construction data is voluminous due to large 70
25 construction industry and discuss the future potential of such volumes of design data, schedules, Enterprise Resource Plan- 71
26 technologies across the multiple domain-specific sub-areas of the ning (ERP) systems, financial data, etc. The diversity of con- 72
construction industry. We also propose open issues and directions
struction data can be observed by noting the various formats
27
73
28 for future work along with potential pitfalls associated with Big
29 Data adoption in the industry. supported in construction applications including DWG (short 74

for drawing), DXF (drawing exchange format), DGN (short for 75

design), RVT (short for Revit), ifcXML (Industry Foundation 76

30 I. I NTRODUCTION Classes XML), ifcOWL (Industry Foundation Classes OWL), 77

DOC/XLS/PPT (Microsoft format), RM/MPG (video format), 78

31 The world is currently inundated with data, with fast advanc- and JPEG (image format). The dynamic nature of construction 79
32 ing technology leading to its steady increase. Today, companies data follows from the streaming nature of data sources such
deal with petabytes (1015 bytes) of data. Google processes
80
33
as Sensors, RFIDs, and BMS (Building Management System).
34 above 24 petabytes of data per day [1], while Facebook gets 81

35 more than 10 million photos per hour [1]. The glut of data Utilising this data to optimise construction operations is the 82

36 increased in 2012 is approximately 2.5 quintillion (1018 ) bytes next frontier of innovation in the industry. 83

37 per day [2]. This data growth brings significant opportunities [Fig. 1 about here.] 84
38 to scientists for identifying useful insights and knowledge.
39 Arguably, the accessibility of data can improve the status To understand the subtleties of Big Data, we need to 85

40 quo in various fields by strengthening existing statistical and disambiguate between two of its complementary aspects: Big 86

41 algorithmic methods [3], or by even making them redundant Data Engineering (BDE) and Big Data Analytics (BDA). The 87

42 [4]. domain of BDE is primarily concerned with supporting the 88

43 The construction industry is not an exception to the per- relevant data storage and processing activities, needed for 89

44 vasive digital revolution. The industry is dealing with sig- analytics [9]. BDE encompasses technology stacks such as 90

45 nificant data arising from diverse disciplines throughout the Hadoop and Berkeley Data Analytics Stack (BDAS). Big Data 91

46 life cycle of a facility. Building Information Modelling (BIM) Analytics (BDA), the second integral aspect, relates to the tasks 92

47 is envisioned to capture multi-dimensional CAD information responsible for extracting the knowledge to drive decision- 93

48 systematically for supporting multidisciplinary collaboration making [9]. BDA is mostly concerned with the principles, 94

49 among the stakeholders [5]. BIM data is typically 3D ge- processes, and techniques to understand the Big Data. The 95

50 ometric encoded, compute intensive (graphics and Boolean essence of BDA is to discover the latent patterns buried 96
inside Big Data and derive useful insights therefrom [10]. 97
These insights have the capability to transform the future 98

of many industries through data-driven decision-making. This 99