Big Data Adoption in the Chinese Construction Industry: Status Quo, Drivers, Challenges, and Strategies

Abstract

Under the influence of pervasive digital revolution, the accessibility and analysis of ‘big data’ can provide useful insights and help various industries evolve. Despite the popularity of big data, the construction industry is lagging behind other industries in adopting big data technologies. This paper fills the knowledge gap by examining the status quo of big data adoption in companies with different sizes and roles, as well as that in projects with different types, and ascertaining the drivers for and challenges in adopting big data. This paper employed a structured questionnaire survey and statistical analyses to investigate the significance of factors influencing the drivers, challenges, and enhancement strategies of big data adoption, and validated the results with post-study interviews with construction professionals. The results show that big data adoption in the construction industry is affected by the size of companies and the work experience of their employees. Technology advancement, competitiveness, and government plan and policy initiatives are identified as the top three drivers of big data adoption in the construction sector. Moreover, a lack of appropriate supporting systems, difficulties in data collection, and the shortage of knowledge and experience are found to be the major challenges in big data adoption. Finally, the identified top three strategies for overcoming these challenges and promoting big data adoption are ‘clear organization structure’, ‘government incentives’, and ‘the training of information technology (IT) personnel’. The paper suggests the necessity of creating differentiated strategies for big data adoption for companies with different scales and roles, and helps provide useful insights for policy-makers in promoting big data applications.

1. Introduction

Statista predicted that the value of the big data market would increase significantly in 2021 and 2022, making it one of the most sought-after commodities worldwide, and the total amount of data currently circulating globally undoubtedly proves this [1]. The data growth brought about by this digital transformation creates both opportunities and challenges for business and research. The availability of enormous amounts of data has given business statistics a completely new push into a direction that is not yet sufficiently understood [2]. Data storage and the management of large volumes of data also challenge traditional statistical and algorithmic methods [3,4], thus pushing companies to create innovative techniques to harness the data.

The construction industry is not an exception to the data boom. The industry traditionally collects data through drawings, material supplies, work breakdown structures, specifications, among others, thus dealing with relatively less data compared to the retail or financial sectors [5]. With the advancement of technologies such as sensors and the Internet of Things (IoT), the industry is faced with increasing amounts of data collected from diverse sources throughout the life cycle of facilities [4]. Analyses of ‘big data’ can be potentially utilized to address current construction challenges and improve productivity.

Big data is a broad term with diverse definitions. Some scholars have defined it as a collection of datasets that are relatively vast and complex, potentially causing difficulties for users in processing them using manual database management tools and conventional data-processing applications [6]. Many believe that the concept of ‘big data’ is characterized by three attributes, namely volume, velocity, and variety (also known as the 3Vs), with the ability to capture, manage, and process data rapidly [4,7]. Volume refers to the size of the data. A huge amount of data from diverse sources provides hidden information and concealed patterns [8]. The sources of construction data include not only project-related design, schedule, cost and quality data, but stakeholder-related corporate data, such as enterprise resource-planning systems and financial data. The common use of building information modeling (BIM) supports the easy capture of multi-dimensional geometric and nongeometric encoded building data [9]. Variety refers to the heterogeneous formats of data. The varied formats of construction data include text, graphs, sensors, audio, reports, and more, which result in challenges for data storage, mining, and analysis. Velocity refers to the rate of data generation [10].

Big data adoption in construction projects has become a trend in solving issues of low productivity, since the industry embraces many innovative technologies that rely on enormous volumes of data [5]. Government initiatives have been set up to promote big data uptake and the IoT, such as Smart Nations, which has, since 2014, been promoted by Singapore [11], and an Internet platform for the Chinese construction industry, set up with the aim of transforming society and businesses through digital innovations. However, the construction sector is lagging behind on the integration of big data and construction processes, making relatively slow progress on the adoption of such innovations [5]. Limited numbers of studies have comprehensively examined the status of big data adoption in the construction sector. Construction companies are often unclear about the potential application areas of big data, big data analytical techniques, and the benefits and potential risks of using big data [4]. While numerous studies have explored the utilization of big data in the construction industry, most existing surveys concentrate on various stages of construction within the entire life cycle [3,12,13]. These surveys do not specifically address these stages across different project types, as this paper does. This paper aims to contribute to a better understanding of the adoption of big data in the construction industry in the following ways: by (1) examining the level of big data adoption in different organizations and project types; (2) identifying the drivers for and challenges in implementing big data in construction projects for organizations; and (3) proposing feasible strategies for overcoming these challenges and unleashing the full potential of big data in the construction industry.

2. Research Background

2.1. The Application of Big Data in the Construction Industry

The construction industry is a relatively fragmented sector, based on projects and multiple organization types, with constantly changing one-off partnerships [14]. Construction data are typically voluminous and heterogeneous due to large volumes of design data, costs, schedules, enterprise resource-planning (ERP) systems, etc.

Big data management cycle mainly consists of four layers, including data source, data ingestion, data storage, data analytics, and visualization [15]. The advance and prevalence of technologies such as BIM and unmanned aerial systems enable construction companies to capture increased volumes of data and to evaluate datasets from multiple sources [5]. BIM captures 3D geometric encoded computation-intensive data in diverse proprietary formats throughout the life-cycle stages of facilities. The design data of a three-story building model can reach 50 GB [16]. Smart sensors with geo-tagging capabilities are attached to the building components for the efficient management of construction material on site and to allow Just-in-Time logistics and inventory planning [17].

Examples of big data storage and processing have been witnessed in a variety of construction areas. Big data and ergonomic methods have been applied to deal with strategic transport safety risks, urban public safety emergency management, and injury and fatality statistics [18]. Big data cloud platforms collect data regarding work behaviors on site from surveillance videos and mobile applications, extract semantic information from images, and identify any unsafe worker behavior automatically [19]. Graph-based big data waste analytics architecture was developed to classify and optimize construction waste [3]. A product-level parallel computing framework using augmented MapReduce was introduced into BIM to improve the efficiency of graphic data processing [20]. A cloud-based system was developed to handle the dynamic data of massive BIMs in 3D. Bigtable and Apache Hadoop were utilized by servers to provide mass storage spaces in a distributed manner, which allowed multiple users to concurrently submit and view BIMs online [21].

Big data analytics provide construction stakeholders with real-time, cloud-powered analytics of construction performance, quality, and inherent risks, thus helping to optimize business processes and improve productivity. Owolabi et al. applied big data analytics to predict the completion risks of Public Private Partnership (PPP) projects, using 4294 PPP project samples in Europe [22]. Braun and Borrmann proposed a method for automatically labelling construction images based on the combination of 4D BIMs and photogrammetry [23]. The proposed method can be used in construction process monitoring and site information retrieval. By analyzing data on safety inspection and projects in Singapore, Poh et al. developed a machine learning approach for developing key indicators for site classification in accordance with safety risks in construction projects [24]. Kim and Chi developed a knowledge management system for construction accident cases using natural language processing [25].

Different projects may have different requirements and barriers, which result in different requirements for big data adoption. For example, big data analysis was used by developers to support a building site selection at Brown University in Rhode Island at the project’s initiation stage [5]. In the bidding stage, developers can select the appropriate bidder based on bidders’ historical performance data. During construction, workers’ routines can be optimized by analyzing their geolocation and movement data from their daily activities via wearables [26]. Moreover, big data can be used to optimize facility management by identifying facilities’ use patterns and maintenance cost patterns, and by planning preventive maintenance [27].

Firms of different sizes have different status quos of and readiness for big data adoption. Coleman et al. showed that only 0.2% of small and medium enterprises (SMEs) in the UK adopted big data analytics in their business, which is far slower a rate than that of large enterprises [28]. Despite the existence of government incentives, SMEs face many challenges before leveraging the advantages of big data, such as cost, technical skills, and organizational capability. Large companies, nevertheless, face problems such as the threat of data isolation in the process of big data adoption.

Despite the efforts mentioned above, limited efforts have been made in real-world projects to process massive amounts of data using big data processing and analytics techniques. Even fewer studies have explored big data adoption in construction companies with different roles/sizes.

2.2. Drivers of Big Data Adoption in the Construction Industry

Nine key categories of driving factors for big data adoption in the construction sector have been identified from a systematic literature review, as shown in

Table 1. An overview of the drivers for big data adoption in the construction industry.

Code	Drivers	References
		[3]	[27]	[28]	[11]	[25]	[5]	[29]	[4]	[9]	[30]	[31]
D1	Economic condition	√			√	√	√	√		√		√
D2	Government support and policy initiatives	√	√				√		√		√	√
D3	Technology advancement			√	√	√	√					√
D4	Employment opportunities	√		√	√		√	√
D5	Competitiveness	√	√	√	√	√			√	√	√	√
D6	Sustainable development			√	√			√			√	√
D7	Initiative to reduce reliance on intensive labor	√	√					√	√	√		√
D8	Workplace safety and health improvement		√	√	√	√	√					√
D9	Increase transparency	√	√		√		√				√

.

(1)

Economic condition

Big data has the potential to boost economic development in the construction sector [3]. According to a report published by the Warsaw Institute for Economic Studies, it is estimated that big data helped to increase the gross domestic product of the 28 EU member states by 1.9 percent by 2020 [29]. According to the report released by Global Construction Perspectives, global construction activities will grow by 85% by 2030, boosted by technology shifts such as augmented reality, big data analysis, and market growth [30].

(2)

Government support and policy initiatives

Government legislations and policy initiatives are vital to the development of big data [25]. The Chinese government revisited its legislative and regulatory framework for technological evolution in order to leverage big data analytics. The Personal Data Protection Act and Public Sector Governance Act have been enacted to provide baseline standards for data sharing and protection across the economy [31]. The government also accelerated the development of next-generation digital infrastructure that serves the needs of the digital economy [32]. The government has also encouraged venture capital and industrial investments in smart construction, and has increased support for the Research &Development of critical technologies for smart construction, the development of basic software and hardware, intelligent systems and equipment, and demonstration projects.

(3)

Technology advancement

The advancement of Information and Communication Technology (ICT) helps to decrease the storage costs, increase the network capacity, and improve the analytics tools and availability of high-performance, on-demand computing through the cloud. Technological advancements allow large volumes of data to be captured, stored, and processed more quickly and accurately [27]. For example, Oudjehane and Moeini used data collected from BIM and drones to assist construction progress monitoring, as well as facility management [5]. In the IoT communication and Web 2.0 era, there are abundant potential data sources that have not been fully leveraged, such as sensors and wearable devices.

(4)

Employment opportunities

The adoption of big data is expected to provide new market and job opportunities in the construction sector. Although big data is emerging as a source of competitive edge in many industry sectors, companies have not benefited as much as possible from big data insights. One of the major reasons is that there are not enough employees with rich data analytic skills [33]. The job market for data analytics is growing in popularity and is expected to increase from USD 130 billion in 2016 to over USD 200 billion in 2020 [34].

(5)

Competitiveness

Companies who have adopted big data would have a competitive edge in identifying and reacting to latent development trends promptly [4]. Decision-making in the construction sector has gradually evolved into cloud computing-enabled real-time decision-making. This data-driven decision-making allows companies to optimize resources, minimize construction waste, and improve productivity. Ahmed et al. found that big data helps to improve the decision-making process, not only at the design stage but also at the facility operation and maintenance stage, by, for example, providing large-scale sustainable design solutions for smart cities and managing facilities in terms of predicted maintenance dates [27].

(6)

Sustainable development

Sustainable cities, smart cities, and sustainable buildings/infrastructure are increasingly gaining worldwide prevalence as a promising approach to combat global warming. In recent years, a large part of research in this area has focused on exploiting the potential of advanced technologies, such as ICT and big data computing, in addressing the challenges of sustainable development [30]. For instance, Bibri applied a data-driven approach to identify the key development trends and forms of smart sustainable urbanism [32]. Gupta et al. presented a case of data-driven decision-making in the supply chain network, supporting a circular economy [33]. In addition, data-driven approaches have been used in building energy consumption analyses for various purposes, such as building energy use benchmarking, energy use mapping, pattern recognition, and load prediction [31]. The internet of things (IoT) and related big data applications can play a key role in improving the process of building construction and operation and in realizing sustainability.

(7)

Initiative to reduce reliance on intensive labor

The Chinese government aims to reduce its reliance on foreign workers for low-skilled jobs, especially for sectors such as construction, security, and cleaning. Around 80 per cent of ground construction workers have a high school education or below, and the overall education level is not high. The outbreak of COVID-19 in 2020 also urged the industry to rethink its reliance on intensive labor. The government works closely with industry stakeholders to enhance the uptake of technologies and the recognition of technologists [9]. The adoption of digital innovations, including big data analysis, could improve the efficiency of the construction sector and gradually change stereotypes in the construction sector, such as low-skilled jobs, high safety risks, and disordered construction sites. Using big data may help to reduce the amount of labor required in construction, and subsequently encourage young people to join the industry.

(8)

Workplace safety and health improvement

The construction industry is accused of higher safety hazards and accident rates than in other industries. Embracing smart technologies such as IOT and robot/robotic arms, with the support of big data, has been applied in exemplar projects to minimize the accident rates on construction sites. Real-time information exchange between sensors, smart phones, and databases offers a new approach to safety management. For example, image data collected from surveillance systems and mobile apps reflect workers’ safety behavior. A big data-based cloud platform can store image data from various sources and extract proof of unsafe behaviors automatically from them [23].

(9)

Increase transparency

The construction industry requires better transparency and trustworthy data [35,36,37]. Big data adoption presents an opportunity to increase transparency among construction stakeholders, since data sharing becomes a requisite when adopting big data. Large volumes of real-time data provide transparency across organizational boundaries, facilitate inter-organizational collaboration processes, and help develop inter-organizational trust [33].

2.3. Challenges in Adopting Big Data in the Construction Industry

Despite the opportunities presented by big data, companies need to address many challenges in dealing with big data. For the construction sector, 21 challenges have been identified from a comprehensive review of the literature, as shown in

Table 2. An overview of the challenges of big data adoption in the construction industry.

Code	Challenges	References
		[27]	[38]	[3]	[39]	[40]	[41]	[37]	[7]	[42]	[5]	[43]
TC1	Storage issues for increased amounts of data	√		√		√
TC2	Data transfer time taken from data collection to processing	√				√						√
TC3	Difficulty of processing varied big data in the construction sector						√					√
TC4	Data ingestion and velocity management	√			√	√						√
TC5	Difficulty of real-time data transmission from construction sites due to limited bandwidth and networking infrastructure	√		√
TC6	The life cycle management of and updates to big data infrastructure			√	√			√			√	√
TC7	Improper representation of data collected			√					√		√
TC8	Lack of effective big data analytical frameworks to meet different stakeholders’ needs		√		√	√						√
MC1	Difficulty of data collection due to the fragmented features of the construction sector	√		√	√		√				√
MC2	Relatively low data quality due to the poor information input environment in construction		√	√				√
MC3	Data privacy concerns from data sources/owners		√	√					√	√
MC4	The lack of adequate data security control systems	√		√						√
MC5	Reluctancy of data sharing among various stakeholders					√				√	√
MC6	Data interoperability issues in the construction sector	√				√	√				√
EC1	Extra investment of companies into hardware and skilled IT personnel			√				√	√
EC2	Cost incurred from continuous technology and skill upgrades							√	√		√
OC1	Necessity of organization structure and daily operation workflow changes for big data adoption			√				√	√		√
OC2	Shortage of skilled data specialists in construction companies	√		√				√
OC3	Lack of knowledge absorptive capacity in innovation		√				√			√
OC4	Challenge for construction workers to accept real time data collection			√				√	√		√
OC5	Possible conflicts of interest between stakeholders	√		√		√					√

. These challenges are grouped into four categories, namely technical (TC), management (MC), economic (EC), and organizational challenges (OC).

2.3.1. Technical Challenges

The technical challenges of big data adoption exist in the data collection, storage, process, and analysis stages. First and foremost, collecting real-time data from construction sites may be challenging. For example, a project site near an underdeveloped area might have problems with low bandwidth, due to the unavailability of sophisticated network infrastructure. The continually expanding amounts and the variety of data produced by multiple sources poses storage challenges for construction companies [39,40,43]. Although distributed file systems and Not Only SQL (NoSQL) databases have been used for big data storage (e.g., Lin et al. [16]), there is still a lack of successful implementation of these in the construction sector.

A relatively long data transfer time, from the data collection point to the storage/processing point, is inevitable; this is another challenge for construction companies confined to a low bandwidth and data transfer rate. Poor data transmission between project sites and the centralized big data repository also affects the real-time monitoring of the project site [27]. Aside from this data transmission issue, data ingestion and velocity management can be another challenge, considering the high dimensionality, velocity, and variety of big data. Moreover, in the construction industry, a wide variety of data is required to be accessed, which includes 2D and 3D data, financial data, enterprise resource information, documents, project schedules, climate data, etc. Processing data from multiple sources in raw, semi-structured, unstructured, and rich media formats is another challenge [43]. Although MapReduce and Spark have been used in pilot cases in information systems of the construction sector, these tools rarely process intertwined data in the various components of BIM.

Furthermore, data renewal in the life cycle process poses another challenge for construction companies. The data stored in the system needs to be continuously reviewed, in order to identify obsolete or invalid data. Additionally, improper data representation would reduce the value of the collected data and undermine the effectiveness of data analysis [40]. Therefore, an effective system for big data analysis is required to meet the different needs of stakeholders.

2.3.2. Management Challenges

The construction sector is well known for its fragmented data collecting, storing, and sharing practices. This fragmented collected data, such as null values, misleading values, and non-standardized values, would be challenging for data processing and analytics [3]. The poor information input environments on construction sites also undermine the data quality, which reduces the efficiency and accuracy of data analytics. In addition, the large amount of data generated from diverse data sources also makes the data quality difficult to judge [37]. Data privacy is another concern for construction stakeholders, since the increasing amount of big data increases the possibility of breaching the privacy of individuals [7,38]. Privacy preservation mechanisms need to be developed for the data life cycle process. Moreover, the development of security control systems for big data is required for big data owners, to ensure the information is resilient to unauthorized users and intrusions [4,41,42]. Furthermore, different construction companies capture their own data from different sources and manage data with different software, such as BIM, scheduling software, and ERP applications [27,42], which increases the difficulty of data interoperability.

2.3.3. Economic Challenges

The hidden cost of big data is one of the major concerns of companies in the construction sector. Firstly, companies need to set up data storage and processing centers, obtain software licenses, and hire skilled IT personnel to process big data [3]. The construction industry is considered to be a low-profit-margin business, so the prohibitive costs of investing in big data are a hurdle for companies [4]. The presence of multiple construction stakeholders, who do not necessarily have the same stake in a project, creates another challenge for big data adoption [5]. Secondly, continuous upgrades of technologies and skills are required to run big data, which incurs more costs [7].

2.3.4. Organizational Challenges

Currently companies do not have sufficiently well-organized processes or effective communication to accommodate the optimization of big data adoption. Therefore, organizational structures and daily operations are expected to change to facilitate big data use. The shortage of skilled data specialists poses another challenge for construction companies. According to a report from McKinsey, there could be a shortage of between 140,000 and 190,000 skilled IT workers to utilize the information gathered from big data [37]. In addition, the low-tech construction industry normally has a relatively low knowledge absorptive capacity to internalize external innovations. Furthermore, more data are accumulated about individuals via numerous means, such as sensors and mobile apps, which raises concerns about violations of personal privacy and security. Developing algorithms that randomize personal data to secure privacy is a key research challenge [40]. The multiple stakeholders involved in the delivery of a project might have conflicts of interest with regard to big data adoption, which poses additional challenges for data sharing and analysis [5].

Despite earlier studies investigating the drivers and challenges of big data adoption in the construction sector, few have explored the status quo of big data adoption in companies with different sizes and different roles, and the differences in drivers and challenges of big data adoption in companies with different sizes and roles.

3. Methodology

To achieve the research objectives, this study adopted a mixed-methods research approach, using both qualitative and quantitative methods, including a comprehensive literature review, an industry-wide questionnaire survey, and follow-up interviews with industry professionals (the overview of the research methodology is provided in Figure 1).

3.1. Literature Review

The literature review was carried out to identify the application areas and benefits of big data adoption, as well as the drivers of (showed in Table 1) and challenges of big data adoption (showed in Table 2).

3.2. Industry-Wide Questionnaire Survey

A questionnaire was developed based on the results of the literature review. A pilot study was conducted to test the validity of the questionnaire, the participants of which were three construction industry professionals with more than ten years of work experience and two academics with research focused on big data. The validated and finalized questionnaire consisted of five sections: (1) the profiles of the survey respondents and their companies; (2) respondents’ viewpoints on the status quo of big data adoption, dependent on various roles, organization sizes, and project types; (3) the respondents’ ratings of the significance and likelihood of the drivers of big data adoption; (4) the respondents’ ratings of the significance and likelihood of the challenges of big data adoption; and (5) the respondents’ opinions on feasible strategies for promoting big data adoption in the construction sector and their judgement on the relative effectiveness of each strategy. A five-point Likert scale, with one representing extremely insignificant or very unlikely and five representing most significant or very likely, was used in Section 1, Section 2, Section 3 and Section 4 of the questionnaire to evaluate the importance and likelihood of the drivers and challenges provided. Another five-point Likert scale, with one representing not effective at all and five representing most effective, was used in Section 5 of the questionnaire to evaluate the strategies provided. The questionnaire script is provided as Supplementary Material S1.

The targeted participants in the questionnaire survey were construction industry practitioners in China holding different roles, including developer, architect, contractor, consultant, quantity surveyor, etc. In total, 800 targeted participants were randomly selected from industrial associations, including the National Certified Constructor Registration System, the China Construction Industry Society, and the China Civil Engineering Society. The questionnaires were disseminated between January and June of 2023 via emails and an online survey tool. By the end of June 2023, 90 responses were collected. Two incomplete responses were removed, yielding 88 valid responses and a response rate of 11%. The profiles of the respondents are presented in

Table 3. Profiles of the survey respondents and their associated organizations.

Characteristics		No.	Percentage (%)
Respondents’ profile (Total = 88)
Role	Architect	13	14.8
	Engineer	11	12.5
	Project manager	22	25.0
	Senior manager	15	17.1
	Administrative staff	8	9.1
	Quantity surveyor	13	14.8
	Facility manager	6	6.8
Year of experience in construction projects	Less than 5 years	12	13.6
	5–10 years	25	28.4
	More than 10 years	51	58.0
Year of experience in big data application	1–3 year	45	51.1
	4–6 years	23	26.1
	More than 6 years	20	22.7
Profile of respondents’ organization (total = 88)
Type	Contractor	30	34.1
	Consultant	25	28.4
	Developer	11	12.5
	Quantity surveyor	6	6.8
	Public sector	13	14.8
	Others	3	3.4
Size of organization	Small and medium enterprise	35	39.8
	Large enterprise	22	25.0
	Multinational corporation	17	19.3
	Others (i.e., public sector)	14	15.9
Year of experience in construction projects	Less than 10 years	15	17.0
	10–20 years	22	25.0
	More than 20 years	51	58.0
Project type organizations mainly involved in	Residential buildings	29	33.0
	Institutional and commercial buildings	18	20.5
	Industrial projects	14	15.9
	Infrastructure projects	21	23.9
	Others (churches, hotels etc.)	6	6.8

. Considering the limited number of respondents with work experience of big data application in the surveyed population, this response rate is acceptable [44].

It was shown that 58 percent of respondents had more than ten years of work experience, and nearly 50 percent of respondents had more than three years of experience in big data. Of the surveyed organizations, 58 percent had more than 20 years of experience in construction practices. Of the organizations that participated in the survey, around 40 percent were small and medium enterprises (SMEs) and 25 percent were large enterprises. All of the organizations had experience in construction projects, of which 83 percent had over ten years of experience. The surveyed respondents and organizations covered various stakeholders in the construction process and had delivered a wide diversity of construction projects, which ensures the quality of the collected data and helps to yield a convincing research outcome.

Four statistical methods were applied to analyze the data collected from the survey, using the software IBM Statistical Product and Service Solutions (SPSS) statistics 27 (shown in Figure 1). First, a one-sample Kolmogorov–Smirnov test was conducted to examine whether the sample data came from a normally distributed population [45]. If the p value obtained from the two-sided test was less than the chosen alpha level (0.05, at a confidence interval of 95%, in this paper), it was suggested that the sample came from a population that is not normally distributed. In the case of distribution-free sample data, nonparametric tests were used as an alternative method to parametric tests, such as t-tests and one-way Analysis of Variance (ANOVA).

Second, as a non-parametric equivalent of a one-sample t-test, a one-sided one-sample Wilcoxon signed-rank test was conducted to test whether the mean of a subject was greater than a critical value [46]. Three was chosen as the critical value in this paper. The objective was to identify the key drivers and challenges of big data adoption in the construction industry.

Third, as a non-parametric equivalent of analysis of variance (ANOVA), a Kruskal–Wallis test was conducted to test the potential difference of means between independent groups [47]. Since the collected survey data was categorized based on respondents’ experiences and roles, a Kruskal–Wallis test was used for inter-group comparison, with the following three objectives: (1) to check whether the perspectives on the drivers and challenges of big data adoption were significantly different among respondents with different amounts of experience; (2) to check whether there were significant differences in drivers and challenges among respondents with various roles; and (3) to check whether there were significant differences in perceptions of the identified drivers and challenges among organizations of different types and sizes.

Fourth, one-sample Wilcoxon signed-rank tests were conducted to identify the key strategies that would help to enhance big data adoption.

3.3. Post-Survey Interviews

Post-survey interviews were conducted with five professionals, selected from the survey respondents, to discuss the results in-depth and to validate the findings. The profiles of the interviewees are shown in

Table 4. Profile of the experts who participated in the post-survey interviews.

No.	Type of Organization	Role in the Organization	Years of Experience in the Construction Industry	Years of Experience in Big Data
1	Contractor	Project Manager	16	4
2	Consultant	Senior Manager	25	5
3	Contractor	Engineer	13	3
4	Consultant	Engineer	27	5
5	Developer	Project Manager	20	6

. All of the interviewees had at least 13 years of experience working in the construction industry and at least 3 years of experience working in big data. In the post-survey interviews, the feedback showed that contractors have relatively limited experience in the application of big data, compared to developers and design companies. Therefore, three years of big data application experience for engineers and contractors was considered to be acceptable. The criticality of drivers and challenges of big data adoption was validated by the experts. In addition, the rationale behind different perceptions on big data by different groups was discussed. The strategies to enhance the adoption of big data were also validated.

4. Results of Data Analysis

4.1. Status Quo of Big Data Adoption in the Construction Sector

The first section of the questionnaire survey investigated the number of construction projects that had adopted big data over the past five years, the areas of application, and the perceived benefits of big data. The results are presented in

Table 5. Profiles of projects which adopted big data.

Project Type	No. of Projects over Past Five Years	Percentage in Total (%)	No. of Projects Adopted Big Data	Percentage in Each Project Type (%)
Residential buildings	1207	25.9	53	4.4
Institution and commercial buildings	1003	21.5	73	7.3
Industrial	874	18.7	82	9.4
Infrastructure	1001	21.5	66	6.6
Others (church, recreation, etc.)	579	12.4	59	10.2
Total	4664	100	333	7.1

. Of the 88 respondent organizations, there were, in total, 4664 projects undertaken by these organizations, of which 333 projects adopted big data techniques, presenting a 7.1% rate of big data adoption. Of the five categories of projects, industrial projects, institutional and commercial buildings, and other project types had relatively high big data adoption rates. Comparatively, residential building and infrastructure projects had lower rates of big data adoption (4.4% and 6.6% respectively).

Table 6. Application areas of and benefits from big data adoption in various construction project types.

Project Type	Residential		Institution and Commercial		Industrial		Infrastructure		Others		Total		χ2 Test
Areas of big data adoption	Frequency	Rank	Frequency	Rank	Frequency	Rank	Frequency	Rank	Frequency	Rank	Frequency	Rank
Assess feasibility of projects	1	4	3	5	3	5	4	4	1	3	12	5	0.263
Improve relationships with clients	3	3	4	3	5	3	3	7	2	1	17	3	0.241
Build stakeholder relationships	1	4	2	6	3	5	2	9	1	3	9	8	0.265
Optimize project design patterns	4	1	6	1	5	2	7	2	1	3	23	2	0.220
Optimize business decisions	0	10	2	6	2	10	1	10	0	6	5	10	0.265
Facilitate construction document management	1	4	2	6	3	5	4	4	0	6	10	6	0.220
Facilitate equipment and asset management	0	9	2	6	3	5	3	7	0	6	8	9	0.265
Reduce safety hazard and risks	1	4	2	6	3	5	4	4	0	6	10	6	0.220
Optimize site management	1	4	4	3	4	4	5	3	0	6	14	4	0.241
Improve tendering and bidding	4	1	5	2	8	1	8	1	2	1	27	1	0.241
Benefits from big data adoption
Improve productivity	76	1	75	1	77	1	76	1	76	1	380	1	0.265
Improve quality	71	3	71	3	69	3	70	3	68	3	349	3	0.241
Increase resource and energy efficiency	74	2	74	2	76	2	76	1	73	2	373	2	0.265
Reduce labor	60	4	61	4	59	4	61	4	60	4	301	4	0.265
Save costs	60	4	58	5	58	5	59	5	59	5	294	5	0.265

presents the application areas and benefits of big data in various project types. The frequency and ranking of these areas/benefits are also presented. It is shown that ‘improve tendering and bidding’ and ‘optimizing project design patterns’ were the top two areas of big data application. Tender price evaluation was also one of the most important areas of big data application. The results of the chi-squared test show that the p-values obtained were above the significance level of 0.05, which suggests that there was no significant correlation between project types and big data application areas.

Likewise, the perceived benefits of big data application included improvements in construction productivity, project quality, resource and energy efficiency, and the reduction in labor and costs. There were no significant differences in rankings of the perceived benefits for various project types. The results of the chi-squared test show that there was no significant correlation between project types and perceived benefits. ‘Improve productivity’ and ‘increase resource and energy efficiency’ were ranked as the top two benefits of big data application by construction practitioners. Comparatively, ‘reduce labor’ and ‘reduce cost’ received lower ranks.

4.2. Drivers of Big Data Adoption in the Construction Sector

Table 7. Data analysis results for the drivers of big data adoption in different types of organizations and respondents.

Code	Mean	Rank	p Value (Kolmogorov–Smirnov Test)	p-Value (Wilcoxon Signed-Rank)	p-Value for Inter-Group Comparisons
					Organization Nature	Organization Size	Organizations’ Years of Experience	Role of Respondents	Respondents’ Years of Experience	Respondents’ Experience in Big Data
D1	3.60	7	0.00 *	0.00 **	0.33	0.56	0.18	0.40	0.52	0.93
D2	3.80	3	0.00 *	0.00 **	0.05	0.62	0.06	0.73	0.26	0.52
D3	4.00	1	0.00 *	0.00 **	0.09	0.02 ***	0.97	0.15	0.19	0.04 ***
D4	3.40	9	0.00 *	0.00 **	0.02 ***	0.25	0.41	0.19	0.80	0.38
D5	3.83	2	0.00 *	0.00 **	0.42	0.58	0.12	0.07	0.13	0.69
D6	3.78	4	0.00 *	0.00 **	0.89	0.66	0.96	0.92	0.74	0.77
D7	3.58	8	0.00 *	0.00 **	0.33	0.25	0.21	0.24	0.89	0.66
D8	3.78	4	0.00 *	0.00 **	0.46	0.05	0.63	0.13	0.62	0.76
D9	3.63	6	0.00 *	0.00 **	0.30	0.77	0.71	0.05	0.41	0.48

Note: * The Kolmogorov–Smirnov test was significant at the significance level of 0.05; ** the one-sample Wilcoxon signed-rank test was significant at the significance level of 0.05; *** the Kruskal–Wallis test was significant at the significance level of 0.05, suggesting the drivers were assessed differently between groups.

presents the analysis results of the drivers for big data adoption in different organizations and project types. The p-values of D1–D9, obtained from the Kolmogorov–Smirnov (K-S) test, were below the alpha value of 0.05 at a confidence interval of 95%, which indicated that the sample data had a non-normal distribution. Therefore, a non-parametric Wilcoxon signed-rank test was used to examine the significance of these drivers. The means of D1–D9 were above three. Moreover, the p-values of D1–D9 were also below 0.05, which indicates that all the drivers examined were significant for big data adoption in the construction sector. The top five drivers were technology advancements (D3), competitiveness (D5), government support and policy initiatives (D2), sustainable development (D6), and workplace safety and health improvements (D8).

The results of the Kruskal–Wallis test identified several drivers that were perceived differently in different organizations and respondent groups. Significant differences were found in the perceived significance of ‘employment opportunities’ (D4) between organizations with different natures. The mean of D4’s significance in quantity survey firms and companies in the public sector was high as 4, while its mean in development firms was 3.08.

Significant statistical differences existed in the perceived significance of ‘technology advancement’ (D3) between organizations of different sizes. D3 was perceived as having the highest significance in multi-national companies, with a mean of 4.53. It was perceived as having comparatively low significance in SMEs, with a mean of 3.74. The p-value of the Kruskal–Wallis test for D3 was 0.02, which was statistically significant in the inter-group comparisons. Moreover, D3 was perceived significantly differently by respondents with different amounts of experience in big data, with a p-value of 0.04. The results in Table 7 show that the amount of experience of organizations and respondents in the construction industry and the role of respondents make no significant difference to the perceived importance of drivers for big data adoption.

4.3. Challenges of Big Data Adoption in the Construction Sector

Regarding the challenges of adopting big data in the construction sector, the impact and likelihood of occurrence of each challenge factor were investigated using the questionnaire survey. The significance of the challenge factors was measured by multiplying the impact of each factor with its likelihood of occurrence. The results in

Table 8. The impact and likelihood of the challenges for big data adoption.

Code	Impact				Likelihood				Significance
	Mean	Rank	p Value (K-S Test)	p-Value (Wilcoxon Signed-Rank)	Mean	Rank	p Value (K-S Test)	p-Value (Wilcoxon Signed-Rank)	Mean	Rank
TC1	3.83	8	0.00 *	0.00 **	3.38	18	0.00 *	0.00 **	12.95	14
TC2	3.38	18	0.00 *	0.00 **	3.20	20	0.00 *	0.00 **	10.82	20
TC3	3.20	20	0.00 *	0.00 **	3.07	21	0.00 *	0.00 **	9.82	21
TC4	3.07	21	0.00 *	0.00 **	3.95	5	0.00 *	0.00 **	12.13	18
TC5	3.95	5	0.00 *	0.00 **	3.68	13	0.00 *	0.00 **	14.54	9
TC6	3.68	14	0.00 *	0.00 **	3.53	16	0.00 *	0.00 **	12.99	13
TC7	3.53	16	0.00 *	0.00 **	3.35	19	0.00 *	0.00 **	11.83	19
TC8	3.35	19	0.00 *	0.00 **	3.70	12	0.00 *	0.00 **	12.40	17
MC1	3.70	13	0.00 *	0.00 **	3.40	17	0.00 *	0.00 **	12.58	16
MC2	3.40	17	0.00 *	0.00 **	3.74	8	0.00 *	0.00 **	12.72	15
MC3	3.74	9	0.00 *	0.00 **	3.56	15	0.00 *	0.00 **	13.31	12
MC4	3.56	15	0.00 *	0.00 **	3.98	4	0.00 *	0.00 **	14.17	10
MC5	3.98	4	0.00 *	0.00 **	3.72	11	0.00 *	0.00 **	14.81	7
MC6	3.72	12	0.00 *	0.00 **	4.23	1	0.00 *	0.00 **	15.74	5
EC1	4.23	1	0.00 *	0.00 **	3.73	10	0.00 *	0.00 **	15.78	4
EC2	3.73	11	0.00 *	0.00 **	4.16	2	0.00 *	0.00 **	15.52	6
OC1	4.16	2	0.00 *	0.00 **	3.90	7	0.00 *	0.00 **	16.22	1
OC2	3.90	7	0.00 *	0.00 **	4.05	3	0.00 *	0.00 **	15.80	3
OC3	4.05	3	0.00 *	0.00 **	3.95	5	0.00 *	0.00 **	16.00	2
OC4	3.95	5	0.00 *	0.00 **	3.74	8	0.00 *	0.00 **	14.77	8
OC5	3.74	9	0.00 *	0.00 **	3.67	14	0.00 *	0.00 **	13.73	11

Note: * The Kolmogorov–Smirnov test was significant at the significance level of 0.05; ** the one-sample Wilcoxon signed-rank test was significant at the significance level of 0.0.

show that the p-values of the 21 challenge factors in the Kolmogorov–Smirnov test were below the alpha value of 0.05 at a confidence interval of 95%, which indicated that the sample data had a non-normal distribution. The results of the Wilcoxon signed-rank test showed that all 21 challenge factors were significant. The top five challenges with the most impact were EC1, OC1, OC3, MC5, TC5, and OC4. The top five challenges with the highest likelihood of occurrence were MC6, EC2, OC2, MC4, TC5, and OC3. Ranked by overall significance, the top five challenges were OC1, OC3, OC2, EC1, and MC5.

As shown in

Table 9. Intergroup comparison results of challenges of big data adoption in different types of organizations and respondents.

Code	p-Value for Inter-Group Comparisons (Impact)						p-Value for Inter-Group Comparisons (Likelihood)
	Organization Nature	Organization Size	Organizations’ Years of Experience	Role of Respondents	Respondents’ Years of Experience	Respondents’ Experience in Big Data	Organization Nature	Organization Size	Organizations’ Years of Experience	Role of Respondents	Respondents’ Years of Experience	Respondents’ Experience in Big Data
TC1	0.10	0.45	0.45	0.44	0.15	0.01 *	0.08	0.50	0.58	0.44	0.55	0.34
TC2	0.78	0.06	0.53	0.42	0.53	0.36	0.08	0.06	0.18	0.42	0.18	0.25
TC3	0.32	0.47	0.92	0.53	0.47	0.83	0.65	0.49	0.84	0.53	0.34	0.80
TC4	0.14	0.94	0.76	0.50	0.90	0.96	0.04 *	0.87	0.61	0.50	0.61	0.17
TC5	0.23	0.16	0.46	0.19	0.09	0.09	0.59	0.74	0.65	0.19	0.12	0.57
TC6	0.40	0.58	0.62	0.94	0.02 *	0.27	0.45	0.29	0.99	0.94	0.20	0.68
TC7	0.28	0.21	0.40	0.87	0.25	0.09	0.06	0.26	0.26	0.87	0.96	0.39
TC8	0.17	0.38	0.61	0.16	0.79	0.37	0.22	0.23	0.25	0.16	0.35	0.17
MC1	0.08	0.35	0.13	0.14	0.80	0.91	0.19	0.73	0.58	0.14	0.38	0.18
MC2	0.05 *	0.75	0.09	0.13	0.67	0.15	0.03 *	0.97	0.50	0.13	0.51	0.15
MC3	0.55	0.23	0.63	0.83	0.08	0.33	0.16	0.92	0.37	0.83	0.07	0.45
MC4	0.16	0.71	0.17	0.29	0.99	0.39	0.19	0.91	0.64	0.29	0.84	0.23
MC5	0.09	0.09	0.17	0.99	0.17	0.20	0.32	0.20	0.09	0.99	0.14	0.34
MC6	0.11	0.15	0.22	0.73	0.92	0.15	0.49	0.43	0.10	0.73	0.64	0.95
EC1	0.20	0.19	0.18	0.06	0.78	0.14	0.07	0.93	0.70	0.06	0.95	0.89
EC2	0.02 *	0.16	0.20	0.42	0.91	0.48	0.02 *	0.40	0.59	0.42	0.88	0.94
OC1	0.52	0.36	0.71	0.89	0.63	0.99	0.23	0.10	0.75	0.89	0.83	0.56
OC2	0.45	0.33	0.29	0.83	0.70	0.80	0.04 *	0.67	0.15	0.83	0.45	0.44
OC3	0.01 *	0.63	0.70	0.11	0.98	0.92	0.08	0.48	0.81	0.11	0.80	0.34
OC4	0.25	0.91	0.67	0.10	0.79	0.96	0.07	0.13	0.81	0.10	0.70	0.25
OC5	0.17	0.27	0.57	0.72	0.22	0.17	0.65	0.66	0.71	0.72	0.36	0.80

Note: * The Kruskal–Wallis test was significant at the significance level of 0.05, suggesting the impact or likelihood of challenges was assessed differently between groups.

, the impact and likelihood of these 21 challenges on different types of organizations and respondents was determined. Regarding the impact of the 21 challenge factors, the results of the Kruskal–Wallis test showed that there was a statistically significant difference in the medians for ‘low data quality’ (MC2), ‘cost incurred from continuous technology and skill upgrades’ (EC2), and ‘lack of knowledge absorptive capacity in innovation’ (OC3) between organizations with different natures. In addition, a statistically significant difference existed in the medians of ‘storage issue of surging amount of data’ (TC1) between respondent groups with different amounts of experience of big data. Respondents with less than three years of experience of big data assigned more emphasis to the data storage issue, with a mean of 4.03. Contrastingly, it appeared to be the case that data storage techniques were relatively mature among experienced practitioners, and some of them even had a comprehensive system.

For the likelihood of the 21 challenge factors, the results of the Kruskal–Wallis test showed that there were statistically significant differences in the medians between organizations with different natures for the challenge factors ‘data ingestion and velocity management’ (TC4), ‘relatively low data quality due to the poor information input environment in construction’ (MC2), ‘cost incurred from continuous technology and skill upgrades’ (EC2), and ‘shortage of skilled data specialists’ (OC2). The size and experience of organizations and the role and experience of respondents made no statistical differences between organizations with different natures. Compared to other types of organizations, such as developers and surveyors, contractors and consultant companies faced more challenges in ensuring data quality, technology innovations, and skill upgrades. In addition, statistically significant differences existed between the medians of ‘storage issue of surging amount of data’ (TC1) between respondent groups with different amounts of experience of big data. Respondents with less than three years of experience of big data assigned more emphasis to the data storage issue, with a mean of 4.03. The likelihood of occurrence of ‘low data quality’ and ‘cost of technology and skill upgrades’, from the perspective of contractors and consultants, was much higher than that from the other groups. Compared to other groups, quantity survey companies and those in the public sector perceived that ‘data ingestion and velocity management’ had a higher likelihood of occurrence, with a mean of 3.67 and 3.53, respectively. Additionally, companies in the public sector assigned a higher likelihood to the challenge ‘shortage of skilled data specialists’, with a mean of 4.08.

4.4. Strategies for Enhancing Big Data Adoption in the Construction Sector

Table 10. Data analysis results for key strategies to enhance big data adoption in the construction sector.

Code	Strategy	Mean	Rank	p Value (K-S Test)	p-Value (Wilcoxon Signed-Rank)	p-Value for Inter-Group Comparisons
						Organization Nature	Organization Size	Organizations’ Years of Experience	Role of Respondents	Respondents’ Years of Experience	Respondents’ Experience in Big Data
S1	Training of skilled IT personnel	4.13	3	0.00 *	0.00 **	0.43	0.29	0.95	0.38	0.69	0.25
S2	Government incentives	4.35	1	0.00 *	0.00 **	0.03 ***	0.54	0.29	0.57	0.27	0.10
S3	Promote successful case study	3.98	7	0.00 *	0.00 **	0.15	0.95	0.17	0.81	0.11	0.77
S4	Promote big data knowledge	4.08	5	0.00 *	0.00 **	0.19	0.33	0.16	0.29	0.43	0.61
S5	Establish standards for data scientists	3.94	8	0.00 *	0.00 **	0.19	0.23	0.30	0.01 ***	0.93	0.37
S6	Clear data governance structures	4.15	2	0.00 *	0.00 **	0.82	0.40	0.61	0.07	0.09	0.71
S7	Top-down leadership	4.11	4	0.00 *	0.00 **	0.48	0.79	0.50	0.04 ***	0.03	0.93
S8	Establish partnerships	3.98	6	0.00 *	0.00 **	0.07	0.67	0.68	0.12	0.28	0.68

Note: * The Kolmogorov–Smirnov test was significant at the significance level of 0.05; ** the one-sample Wilcoxon signed-rank test was significant at the significance level of 0.0; *** the Kruskal–Wallis test was significant at the significance level of 0.05, suggesting the strategies were assessed differently between groups.

presents the analysis results for the enhancement strategies, S1–S8, for big data adoption. The p-values of the K-S test for the eight strategies were below the alpha value of 0.05, which indicated that the sample data had a non-normal distribution. The results of the Wilcoxon signed-rank test indicated that all the strategies were significant for big data adoption in the construction sector. The top three strategies were ‘government incentives’ (S2), ‘clear data governance structures’ (S6), and ‘training of skilled IT personnel’ (S1).

‘Government incentives’ was perceived as the most important strategy for enhancing big data adoption, with a mean of 4.35. ‘Clear data governance structures’ was ranked as the second most important strategy for promoting big data adoption.

The results of the Kruskal–Wallis test identified several strategies that were perceived differently in organizations with different natures and by respondents with different roles. Significant differences in the medians were found in the perceived significance of ‘government incentives’ (S2) between organizations with different natures. S2 received a higher mean perceived significance from contractors, consultants, and developers than from quantity surveyors and companies in the public sector. This demonstrates the critical need for government incentives to encourage the development of data applications by such practitioners.

Significant statistical differences existed in the perceived significance of ‘establish standards for data scientists’ (S5) and ‘top-down leadership’ (S7) between respondents with different roles. This showed that different roles had stricter requirements for data compliance and standardization. S5 was perceived as more significant by quantity surveyors and engineers, with a mean of 4.67 and 4.53, respectively. The p-value of the Kruskal–Wallis test for S5 was 0.01. Moreover, S7 was perceived significantly differently by respondents with different roles, with a p-value of 0.04. Consultants and developers placed more emphasis on the significance of top-down leadership than other groups, indicating that consultants and developers benefitted from a structured data processing procedure when applying big data. This is the key to ensure the data is handled smoothly. According to the experts’ viewpoints in the post-survey interviews, a data governance structure involves “decision-making, management and accountability related to data in an organization, and … specifies the data ownership and the responsibilities of stakeholders”. A working group for data governance should also be established to liaise between business and data-related technologies, as follows: “The team drives the big data use and ensures data quality for specific areas, should involves experts specialized at both business and IT issues”.

5. Discussion

5.1. Application Status of Big Data in the Construction Sector

According to the survey respondents’ opinion, data standards and specifications in residential building projects are still inadequate, which limits the adoption and promotion of big data technology. Residential construction projects typically have long life cycles, with each phase taking a significant amount of time, from design and construction to delivery. It means that, even if big data technology is introduced, it will take a longer time for its effects to be felt. The data collected from the survey suggested a lack of knowledge about big data in the construction sector. Many respondents misunderstood the definition of big data, especially small contractors, who assumed that any kind of data used during operations is considered to be big data. Despite some practitioners recognizing the importance of big data in fostering innovation in the construction sector, big data is still infrequently used in this discipline.

With the improvement in information technologies, projects’ cost-related data can be recorded systematically in real time, which supports tender price evaluation and decision-making for the cost management of construction projects. An internal database for the electronic tendering and bidding system can provide a more objective and fair pre-assessment for all parties participating in the bidding process [15]. In addition, the increasing data amounts and data sources represent an opportunity to innovate how firms design, such as data-driven design and iterative design. According to respondents’ feedback, the building information collected can be fed back into the early design stage, which allows the design team to understand how occupants interact with the built environment and to tweak and adjust the design. Firms can develop the optimal design by testing hundreds of programmatic and operations inputs from clients, as well as external environmental factors.

Tools such as BIM help make the construction industry more economically and environmentally sustainable, and play a massive role in improving energy efficiency. BIM assists stakeholders in visualizing each stage in the construction process, helps streamline the process, and reduces material wastes.

5.2. Key Drivers of Big Data Adoption in the Construction Sector

Figure 2 presents the overall analysis results of the drivers, challenges, and strategies for big data adoption in the Chinese construction industry. ‘Technology advancement’ is perceived as the most important driver of big data adoption. Technologies such as ICT, cloud computing, BIM, and drones not only enable organizations to gather more data from diverse sources, but allow for improvements in data storage and analysis. ‘Competitiveness’ is ranked as the second top driver in adopting big data. The multiple players in the construction industry drive organizations to keep a competitive advantage and outperform their rivals [37]. Strengthening data innovation maintains the competitiveness of practitioners of the construction industry to a large extent. Comparatively, ‘reduce labor’ and ‘reduce cost’ received lower ranked perceived significance. Many respondents believed that big data had little effect on labor reduction and cost saving, the results of which might be associated with the challenges of big data adoption.

The factor ‘employment opportunities’ (D4) was perceived differently by survey firms, companies in the public sector, and developers. The public sector pays more attention to the allocation of basic social resources such as jobs, which is not the focus of developers. ‘Technology advancement’ (D3) was perceived differently by differently sized companies. Large companies’ production scales tend to be saturated, and they are more urgently in need of advanced technology to support patterns’ innovation and to consolidate their industry position; SMEs are adamant about increasing the number of construction projects to expand their market share, and lack the corresponding financial support to drive technological progress. Moreover, respondents with more experience in big data place more emphasis on technology advancements. The first companies in the area of big data or those with a deeper understanding of the practical applications of big data in the construction industry will cherish the convenience brought by advanced technologies.

5.3. Challenges of Big Data Adoption in the Construction Sector

The top five highest-ranked challenges are OC1, OC3, OC2, EC1, and MC5. The results suggest that, from the viewpoints of respondents, the most significant challenges in adopting big data exist in the organizational aspect, which is consistent with the findings of Hwang et al. [14] and Coleman et al. [28]. The establishment of an appropriate organization structure, a large knowledge absorptive capacity, and data specialists are necessities for a company when adopting big data. Additionally, the extra investment of big data and the reluctancy to participate in data sharing are major concerns for companies in big data adoption.

The results suggest different hindrances to big data adoption in different types of organizations and respondents. Compared to other types of organizations, such as developers and surveyors, contractors and consultant companies may face more challenges in ensuring data quality, technology innovations, and skill upgrades. They serve as project implementers throughout the project implementation cycle. Consultants design the project in accordance with the developer’s requirements and the results of the engineering survey, as well as take on the responsibilities of optimizing the construction plan and analyzing market demand in order to meet business needs and expectations; contractors are in charge of the project’s drawings and design, turning the engineering blueprint into a solid concept. A large amount of data circulation is interspersed in this process, so they will have higher requirements for data quality and technological advancements, accordingly.

5.4. Key Strategies for Enhancing Big Data Adoption

At the current stage of big data application, government incentive still plays a dominating role in promoting big data in the construction industry. Carefully designed policy strategies help construct the social norms of big data usage. The government incentives that promote big data application mainly include positive feedback and negative feedback. Positive feedback includes financial incentives and subsidies for big data and open data policies for the public. Negative feedback refers to regulations and legislative frameworks for data privacy and corporate control through data [48]. The government should strengthen top-level design and optimize the management mode of enterprise big data application [15], build a unified standard system, and accelerate the formulation and promotion of standards for the design and construction, technical application, acceptance assessment, and safety assurance of big data in the construction industry. ‘Clear data governance structures‘ is another necessity in the adoption of big data by companies. Another significant strategy that helps promote big data adoption is the ‘training of skilled IT personnel’. There is a lack of highly skilled IT personnel in the construction industry. Skilled IT team members should be trained to provide technical tools for the business team, and to code for data storage and data models. According to the experts’ viewpoints, “They need to understand how to translate business models into data models and determine rules for cleaning the data. They also need to collaborate with business departments in the organization, understand their data needs”. In addition, IT experts also "should provide metrics to help the business experts evaluate data quality and policy adherence and determine if any changes need to be made”.

6. Conclusions

‘Big data’ has been growing in popularity in recent years, and the application and analysis of big data has been fueling innovation across various industry sectors. Despite the notion of ‘big data’ gaining traction in the construction industry, it is still in the nascent stage of adopting big data technology and lags behind other industries. Limited literature exists on the status quo and influencing factors of big data application in the construction industry. Thus, this study used questionnaire surveys and interviews to gain a better understanding of big data adoption in the industry.

Firstly, the study analyzed the ratio of big data adoption, application areas, and perceived benefits in five different types of projects. The results revealed an overall lack of knowledge about big data in the construction industry. Industrial projects and institutional and commercial building projects have a relatively high big data adoption rate. The main application areas of big data are “improving the tendering and bidding process” and “optimizing project design pattern”. “Improving productivity” and “increasing resource and energy efficiency” are identified as the top two benefits of big data application.

Secondly, the study also identified five major drivers of the big data adoption in the construction sector, including “technology advancement”, “competitiveness”, “government support and policy initiatives”, “sustainable development”, and “workplace safety and health improvement”. The impact and likelihood of occurrence of each challenge factor were also investigated by this study, the results of which show that the most significant challenges in adopting big data still exist in the organizational aspect, despite the inherent technological complexity of big data. Meanwhile, the extra investment necessary for big data adoption and the reluctancy to participate in data sharing are the two major concerns for companies in the decision-making process of big data adoption.

Finally, a list of strategies has been proposed for construction practitioners to improve big data adoption. The top three strategies are “government incentives”, “clear data governance structures”, and “training of skilled IT personnel”.

The findings of this study are valuable, as this study uses first-hand survey data to identify the status quo of big data application in the construction sector and assess the impact and the likelihood of occurrence of the drivers and challenges of big data adoption. In addition, the strategies formulated in the study encourage construction industry practitioners to efficiently apply big data in projects, and to optimize resource allocation from a long-term perspective, thus rejuvenating the construction sector in the information revolution. The study contributes to a systematic big data transformation process involving multiple elements, including but not limited to top-level design, organizational culture construction, the application of technologies, data management, business process optimization, and the digitization of project management. Enterprises need to formulate clear digital transformation strategic planning, and clarify the development goals and directions of transformation and upgrading. Enterprises should strengthen organizational culture construction, deepen enterprise reform, build a high-end digital talent training system, and create a development environment suitable for transformation and upgrading. Relying on BIM, Internet of Things, cloud computing, blockchain, and artificial intelligence, a digital construction innovation platform should be created to open up digital space and physical space, and to build a digital industry internet platform for the different life cycle stages of project management.

Nevertheless, there are some limitations of this study. The sample size of the questionnaire survey and interviews is relatively small, leaving room for future investigation in a wider context. Moreover, the findings of the study are based in the context of China, and care should be taken when generalizing the findings to other contexts. Future research could also investigate the association between the drivers of big data application in the construction sector and establish an evaluation model to test the effectiveness of the proposed strategy for promoting big data adoption.