The four Vs of big data
Data is integral to successfully using any construction technology
Data underpins any plans to use digital technologies in construction and ensuring it is consistent, without bias and clean is essential for success. Increasingly, examples of transformational digital construction projects are highlighted as case studies, such as the use of remote sensors, materials tracking, BIM and digital twins, but what all will have in common is use of data.
Without a strong strategy around the collection and use of data, digital construction will have limited success.
The success of construction projects is typically measured in terms of time, cost and quality. Underpinning this success, a factor that makes the difference on meetings these measurements is having the insight to make timely decisions.
Speaking during the Construction Technology Forum, Mohamed Nagi, Digitization Manager, Analytica Management Solutions (ASGC Group), talked about there being four characteristics of big data that construction companies had to consider if it was to make step changes in the information provided and transform processes. He called them the four Vs of big data:
Very high speeds
Variety of structured and unstructured data
Nagi’s argument is that high volumes of data need to be transferred at very high speeds. These data sets will include both structured and unstructured data (variety), and must be accurate (veracity).
If used correctly, big data can be transformational, he said. “The idea behind big data is that everything we do is digitally traced. Everything you do on a construction site is leaving a digital trace. Large data sets driven from multiple platforms at very high speeds include structured and unstructured data. They should be able to provide insights and patterns, and correlations that improve decision making and optimise a business.”
Structured data is data that sits in a fixed field in a record or file, while unstructured data does not sit within a field. An example of unstructured data could be photos from the construction site; once those image files are tagged, they become structured data. It is the difference between data and turning data into intelligence – actionable information has been structured and has value.
Veracity is important because if the data is not accurate, nor will the information it provides be, but ensuing data is input without bias can be difficult, partly because that bias can often be unintended. Biased data can happen because people make mistakes or they withhold data (knowingly or unknowingly). The end result though is it will influence decision making wrongly, because it is based on the information provided. It means understanding how data is collected – is it observational, sensor-based, collected via technologies such as Bluetooth?
A well-documented example of unintended bias leading to incorrect reporting came from UK government efforts to introduce a track and trace scheme for Covid-19. It developed a mobile phone app to monitor if a person had been in close contact with someone subsequently found have the virus. But it was discovered the app could not distinguish if someone had been in close contact, or on the other side of a wall or door, or in a vehicle, due to it using Bluetooth.
The role that data will play is still evolving and the use for some of it will be unclear, potentially leading to it being ignored. But that data could provide both current and historical information in future and be important for Artificial Intelligence and machine learning.
This will mean asking people to record data that they don’t see the benefit of collecting, particularly difficult in more transient situations. A way around this disconnect is to ensure that the ‘currently useful’ data is used as a tool in real-time decision making, such as integrated into dashboards that work on mobile devices and provide visual information that can undergo simple manipulation.
To provide analytics and information on project progress for instance, it means setting KPIs around what data needs to be collected onsite. That data can include performance measurements of equipment, downtime, people numbers and so on.
“The goal is to develop and provide our teams onsite the tools to help them build more efficiently, [such as] improving the project controls onsite that give real-time measurements of the project progress [or] real-time control of the resources being used,” said Javier Bonilla, Head of the Technology Observatory for Construction at Acciona Construction, during DC Hub’s webinar on 10 technologies disrupting digital construction.
The use of construction technology is rising onsite. Overtime, the introduction of new tools to help decision-making will become faster and easier, because the data is being collected and can be used for current and historical information.
It will also introduce challenges such as security and protection (as the industry becomes more data rich, it becomes more attractive to cybercriminals), and data ownership (is it the client, contractor, or software developer, for instance). These are going to big significant issues in the future that we will look at separately.
This article first appeared on sister site Digital Construction Hub